NTSB/AAR-94/07 - AirDisaster.Com

WASHZNGTON, D.C. 28594
STALL AND LOSS OF CONTROL ON FlMAL APPROACH
ATLANTCC COAST ARLINIES, INC./
UNIITED EXPRESS FLIGHT 6291
JETSTREAM 46 01, N304UE
CQLUMWS, OHiO
JANUARY 7; 1994

P394-910409
ATLANTIC COAST AIRLINES, INC./UNIHED EXPRESS FLIGHT 6294
ETSTREAM 41 01, M903UE
COLUMBUS, OHIO
JANUARY 7,1994
Adopted: October 6,3994
Notation 6338A
Abstract: This report expla!ns the crash of United Express flight 6291, a Jetstream 4i 01
airplzne, whiie CR zqproach to runway 28t at Port 63lumbus internationzl Airpmt,
Coiurnbus, OMo, cn january 7, 994. The safety issues ir: the report focused OR aircraft
belts, iiiining propms ?os Pzrt 135 pi!& that emphaska st&: warming
rocopition and recovery iachniquas, and tha? lead io proficiency in both hrgh sped and
cocpled approaches. Sate!)! rec3mmenda?ions concerning these issues were made to
t3e &derai Aviati~n Administrsticn.
h

CONTEhTS
EXECUTIVE SUMMARY .................................................................
1 .
1.1
I . 2
1.3
FACTUAL rn!ORMAZ'HON
History of the Rig& ..............................................................................
hjuries to Fe'ersons .................................................................................
Damage t3 Xiraft ................................................................................
1.4
1.5
1.5.1
1.5.2
1.6
1.6.1
1.7
1.8
1.9
1-10
1 ..I
1.1 1.1
Other Damage .......................................................................................
Personnei Information ..........................................................................
The Laptajn
" ...........................................................................................
The First 0,s cer ....................................................................................
hAzi Information ...............................................................................
The Autopilot ........................................................................................
Meteoroiogicd Information ...................................................................
-43s to Navigation
. .
................................................................................
co.n~Knlcations ....................................................................................
Aerohme Inform~tion .........................................................................
Right Recorders ....................................................................................
Coc'qit Voice RecorCer ........................................................................
1.11.2
i.12
I.i3
Flight Data Recor~er .............................................................................
Wreckase md lmpaci IrJormation .........................................................
hledicaf uld Pathologird Infomation ...................................................
1.14
1.15
1.16
1.15.1
1.16.2
1.16.3
1.16.3.1
Em .......................................................................................................
SurVivaI Aspects ....................................................................................
Tests and Research ................................................................................
Flightpath Reconstnrc3tion ......................................................................
Fiigfit Tests ...........................................................................................
System Teardom and Examination .......................................................
Ice and bin Protection System .............................................................
1.16.3.2 Stail kVamkxg Systems ...........................................................................
1 . i 6.3.3 Ground Proximity W b g System (GPWS) .........................................
1.17 Addition& Infomation ..........................................................................
I .1 7.1 Corporate History and Organization ......................................................
1.17.2 BA-SIPGTTraiing .................................................................................
- 1.1 i . 2.1 Gramd Training ....................................................................................
..
1.172.2 Right Tra-mg .......................................................................................
1.1'7.3 Altitude and Airspeed Awareness ..........................................................
I .I 7.4 FAA Survei!tmce ..................................................................................
V
1
6
6
6
7
7
9
11
11
13
17
17
17
18
18
18
19
21
22
23
25
25
32
34
34
36
37
38
33
41
41
42
43
45

R
2 .
2.1
2.2
2.3
2.4
2.5
2.4
2.7
2.8
3 .
3- 1
3.2
4 .
5 .
ANALYSIS
General .................................................................................................
Flightcrew and Aircraft Performance .....................................................
Pilot Trainiig and Experience ................................................................
"be Company ........................................................................................
FAA Surveil!arxe ..................................................................................
Comctive Actians - ................................................................................
.. TT
Occupant Safeiy bel[ usage ..................................................................
Additional Ldormation .........................................................................
CoNeLusroNs
Fmdmgs ................................................................................................
Probable Cause .....................................................................................
RECOMMEFJBATIONS ....................................................................
APFENDMES
Appendix A-hvestigation and Hearing .................................................
Append= €+-Cockpit Voice Recorder Transcript ..................................
Appmdix G-Safety R?comenda;ions .................................................
iv
50
51
63
64
68
48
69
78
71
73
74
77
78
103

EXECUTIVE SU~WPvfARY
On January 7, 1994, about 2321 eastern standard time, a Je.tstream
410i, regismtion N?04LI, operated by Atlantic Coast Airlines, Sterling, Virginia,
and doing business ES United Express flight 6291, craskd 1.2 nautical rniies east of
runway 2% at Port Columbus International Airport, Columbus, Ohio. The airplane
was being operated as a regulariy scheduled commcter flight under 14 Code of
Federa! Regdatinns, Part 135, &om Washington Duks International Aiqortt
Chantilly, Virgkia, to Coluxnbus, Ohio. The flight had been cleared for an
instrumen: landing system approach ta runway 28L and was in contact with the
!ma1 tower controller when i$ crashed into a storage warehouse. The pilot, copilot,
fiig5t atteedani, and two passengers were fatally in; -ed. Two of the other three
passerigers received minor injmies, while the third was not iajurrd. The airplane
was destroyed. Instrument meteorological conditions prevailed at the time, and the
airplane was on an inswment flight rules flight plar?.
The National Transportation Safety Board determines the probable
causes of this accident to be:
(1) An aerodynamic stail that occurred when the flightcrew allowed
the airspeed to decay to s’dl speed following a very poorly planned
azd executed approach characterized by an absence of procedural
discipline;
(2) Improper pilot response to the sta!t warning, including fahre to
advvlce &e power levers to maximum, and inappropriately raising
the fiaps;
(3) Flightcrew inexperiencz in “giass cockpit” automated aircraft,
aircraft type, and in seat position, a situation exacerbated by a side
ietter of qeemeni between the company and its pilots; and
(A) m-e r~qznny’s f&Are sm:;i& &ywat.- rtshilized approach
1----
--“*I* ~
criteria, and fhhe Federal Aviation Administration’s failure to requirr
such criteria.
Mem’ber Vogt concluded that the last fxtor was contributory but not
cairsa! EO rhe accident. Additionally, for the following two factors, Chairman Hail
aqd Member kati‘mr concluded that L\ey were causal to the acd-nt, while
E‘

Members Vogt md Hammerschmidt concluded that they were contributory to the
accident:
(5) The company’s failure to provide adequate crew resource
management mining, and the FAA’s failure to quire such training;
and
(6) The unavailability of suitable training simulators th2t precluded
fully effective flightcrew training.
Safety issues discussed in the report includp aircraft safety belts, and
training programs for Part 135 pilots that place more emphasis on stall warning
recognition and recovery techniques, and that train pilots to proficiency for both
high speed approach profiles and coupled approach profiles. Safety
recommendations concerning these issues were made to the Federal Aviatim
Administration. Also, the Safety Board reiterated safety recommendations to the
Federal Aviation Administration concerning stabi!ized approaches and aircraft
safety belts.
vi

NATIONAL TRANSPORTATION S AF"Y BOARD
WASHIEGTON, D.C. 20594
AWCRAFT ACCIDEW REPORT
STALL AiW LOSS OF CONTROL ON FPNAL APPROACH
ATLANTIC COAST AfRLINJS, INC./UMTED EXPRESS FLIGHT 6291
JETSTREAM 4101, N304UE
COLUMBUS, OHIO
JANUARY 7,1994
1. FACTUAL INFORMATION
On January 7, 1994, about 2321 eastern standar:! time @a),1 a
Jetstream 4101 (J4101),2 registration N3WUE, operated by Atlantic Coast Airlines
(ACA), Sterling, Virginia, doing business 2s (d/o/a) United Express flight 6291, call
sign B1ce Ridge flight 29!? crashcd 1.2 nautical miles east. of runway 2% at Port
CoImbus International Airport (CREI), Coiumbus, Ohio. The airplane was being
operated as a regularly scheduled commuter flight under 14 Code of Federal
Regulations (CFR), Part 135, from Washington Dulles International Airport (IAD),
chantilly, Virginia, to Columbus, Ohio. Tbe flight had been cleared for an
instmmec: L:2zg system (ES) approach to runway 281, and was in contact with
the local towe; controller when it crashed iaio a storage warehouse. The pilot,
copiioi, flight attendant and two passengers were fatally injured. Two of the other
three passengers received minrJr injuries, while the third was not injured. The
airplane was destroyed. Instrument meteorological conditions prevailed at the time,
and the flight was on an instrument fiight d es (n;R) flight plan.
'AH thcs xe atern standard he (EST) based on the 24-hour clock, unless otherwise
indicated.
h e J-3101 ray Ix refed to as a 1-4100, J-4101, BA-4100, or BA-4101 in this report since
the mmufzcwer has changed nams from British Aerospace (BAe) to 3etstream Aircraf: Limited
The "I" or "0" suffix at the ezd of the modeI number refers to an airplane flown in the United
States or the United Kingdom, respectively.
'UP>?& Express Flight 6291 nsed the call sign of Blue Ridge 291 for air traffic control
com,ufiications.

2
Company records indicated thzt the accident airplane h ved at UD
from CMH at 2105. The pennimting crew stated that the airplane had flown six
flights that day for E 2 kours of block ?he. tight icing was experienced on all six
flights. Tlie terminatiig crew reported that the airplane had pedomed satisfactorily.
The anti-ice and deice systems were tested and operated properly several times
during the day. No discrepancies we= recorded in the airplane logbook.
The captain of flight 299 requested &at the airplane be loaded with
4*l@@ norm% of fuel, and he selected Dayton, Ohio DAY), as the aItemate
destinakon. He proceeded to th~ aippIane whese he met bnefly with the first officer
who had flown the inbound fiighi frcm ChH.
After the passenger boarding was completed, the crew performed
weight mnd balance calculations. The flightcrew moved four passengers fran?.
Section A to seats in Section C to achieve the proper balance for the airplane. This
was refiected on the completed weight and balance caiculations for the accident
flight
The airplane departed the gate at 2158 (as reported by ACA flight
following) with five passengers on board, Tne flight was planned to cruise at an
altitude of 14,000 feet above mean sea level (mi) wit% an en route fligh: time of
1 hour and 30 mbutes.
Whiie en route, at 225930, the radar controller at the Indianapolis Air
Roue Traffic Control enter (ARTCC), staled, “Blue Ridge 291, be advised, ah,
just had a repart of some icing at 14,000, 10 o‘clock to you and, ah, about 25 -
30miies.” The captain, who was flying the airplane, replied, “You said that was
some light rime?” The controller then transmitted, “404CK, what kind of icing were
you getcing?” The pilot of the aircraft responded, “h.oderate, moderate rime on up
to 14,320 and we’re, ah, we’re in the clear, ~5: b the clear above us up here at
15,oOO.” After receivilg the report and thanking the pilot, the controller relayed,
“Blue Ridge 291, he said it was moderate rime iCig up $9 14,GOO.” At 2302:28, the
fli&tcrew of Blue Ridge 291 inquired, “And hdjmapolis Center, Eke Ridge 291,

3
can we get, ah, 15,009 for a littie while?” The request was granted. At 23%:42,
the radar controller transmitted, “And Blue Ridge 291, pilots discretion maintain
1 1,oeQ,” to which the flightcrew xepeated the Clem=, acknowledging the call.
About 2310, flight 291 contacted CMH approach control and advised
the controller that they were descending though 13,200 feet for I1,OOO feet and thal
they had Antomtic Terminal Information Service (ATIS) information Alpha. ATIS
information Alpha was as follows:
Measured 1,100 overcast, visibiiity 6, light snow, fog, temperature
23, dew point 22, winds 330 at 4, altimeter 29.97, Ei Runway
28 Left approach in use, also landing runway 28 right.
Right 291 was assigned a 285-degree heading to intercept the IIS for
---..-> =:mar‘ --- 38T. at C-MH and was cleaEd to descezd to !O,ocX, feet At 2315, fiight
291 was advised of the updated weather report at CMH (AITS Bravo), which the
crew a&noivIedged. It read as foIlows:
Special weather 0410 Zulu [2310] measured ceiling 8W overcast,
visibility 2 !,2,
light snow, fog, wind 3W EX 4, altimeter 29.97.
At 231628, flight 291 was advised of their position, 10 miles from
SLJ,’ to maintain 3,000 until established os the :ocalizer, anand was cleeared for
the I U approach to mwzy 28L ai CMH. (See Eguxe 1). The flight zcknowkdged
the clearance. About 1 mkute lakr, air W e control (AX) instructed flight 291 to
reduce its speed to 170 bots and to contact the CMH tower controller.
At 231,758, sounds shilar to a reductior? in propelIer/engine noise
amplik& is noted on the cockpit voice recorder (CVR) transcript. The full CYR
tmscript, including all AT6 tmmmissicns within that period of time, is included in
appendix B.
At 2318:20, the crew contacted the CW3 local tower contrdler.
controller cleared rl,ght 291 to land on runway 28L. The last transmission to ATC
received from the airplane before the accident was the acknowledgment of tint
dearance.
%ne iridal approach fix (lA3 for the IL3 approach to runway 28L at CMH.
.a

A.
4
I 1230'*
1 . Not for Navigational Purposes
1 MISSED APP~KW Climb to 2700' direct CB LOM and hold.
Figure 1 .--CMH approach chart.

portion of the flight were:
23 18:44
2318:46
231853
2319:14
231930
2319:36.8
2320:01.3
2320:02.3
2320:08.5
232025.6
232031.6
232036.1
232035.1
232039.8
2320.41.1
232041.6
232042.7
232044.5
232a46.2
2320:45.6
232047.2
232Q48.1
232049.5
232050.2
2320:58.8
232051
232052.3
First Officer
Captain
First Officer
Caplain
Captain
First Officer
captain
Captain
First 3fficer
First officer
Captain
First Officer
Captain
First Officer
Captab
First Officer
First Officer
Captain
captain
First Officer
ref is 112, I gotta plug that (too)
1 did it for you
here comes the glideslope
sounds similar to altitude or gear
warning alerts
and wefe maker inbound
don't forger to give me my caiis i,X4 is
DH
a thousand above
okay flaps nine
geaI down
flaps rifteen landing gear down three
green
condition levers a hund- condition
levers a hundred percent
okay give me a hunk4 pem2t please
a hundred percent flows at three
sounds of increase in propeller/engine
Bg*
three
yaw damper
and autopilot to go don't touch
don't touch
holding on the yaw damper
s~m& similar KI bha2. Jf a stick shaker
Start
sound of seven tones similar to that af
autopilot disconnect alert
Tony
sounds similar to a stick shaker stop
what did you do?
I didn't do nothing
sound similar to lhat of stick shaker
St2rtS
sounds similar to that of an increase in
prop/engine noise amplitude

232052.5
232053.7
232053.7
232054.0
232054.3
232055.3
232357.5
6
captain g h e flaps up
sow.ds sk&r
stop
to that d stick shaker
CaPQiXl no no hold it
the GPWS transmits '%I!"
soinds similar to that of stick shaker
starting again and continuing to the end
of zecording
captain gimme flaps up
sc~~~ds rid= tn fht of change in or
addition io stick shaker
captair, whoa
sound of impact
The crash occurred about 2321 during the hours of darkness. The
airplane came to rest about 1.2 miles east of the threshold of Cms runway 28L at
39" 59' 31.8" noah latihde and 8 2 O 50' 49.8" west !ongitude. The accident site
elevation was 866 feet msl.
1.2 Injuries to Persons
Injuries Flightcrew Cabincrew Pzssenpers Other
Fatal 2 1 2 0 5
Serious 0 0 0 0 0
Minor
None
Total
0
- 0
2
0
- 0
:
2
- 1
5
0
- 0
0
2
1
8
-
1.3 Damage to Aircraft
The airplane was partially destroyed by %e impact with trees and a
concrete block structure. It was then consumed by a postcrash fm. The airplane's
value was estimated at $7 million.
1.4 Othei Damage
Numerous trees, the highest of which was approximately 60 feet tall,
were destroyed. A storage warehouse, built in 1989, was destroyed when &e

1.3 Personce! Information
1.5.1 The Captain
7
The czptai?, age 35, had been hired by ACA on April 13, 1992, as a
first officer on the Mstreeam 3201 (3-3201). He completed transition training and
upgraded to captain in the J-4101 on October 26,1993. He held an airline transport
piiot (ATP) cerlificate with mtings and limitations for airplane multiengine land,
BA-4100, and commercial pilot privileges for shgle-ertgine land airplanes. He also
possessed a flight instfictor certificate with ratings and limitations for airplane
single and multien,oine, instrument airplane. His total pilot &me was 3,660.4 hours,
oi which 1,373.4 horn were &&0pr00 airplmes. He had acc-mulared
191.9 hours in the J-4101, of which 150.8 was log_red ifs he pilot fligfft time ipiiot-
bncommand (PIC) rime), 39.1 hours night experience, md 23.5 hours instrument
experience. The captah possessed a first class medical erlificztte issued on
September 20,1993. It contained na limitations.
The captain earned a Bachelor's degree in urban systems and had
previously been employed as a computer pmgmrmer. He began his a-:iation career
%I l98.5, and had worked as a rlight instructor in 1987. He worked periodically as a
%SM 2% charter pilot in th~ northeastern United States €or about 4 years until he
was hired by ACA and entered Jetstree 3200 fmt officer training. He completed
trainicg but failed kk initial second-&command (SIC) 14 CFR Part 135.293
sirnulator check on May 7, 1992, as a mult of dif5dSeiies +*th Lmtmnent
" -..-... -* --I_
-------&-.- .-A
a~pvaa.lrw auu il"l-r_r
Lr.IA&- -.---A*- p-cuu'cs. A k n 2 ~ A hnlltr dAitign~1
1 UIIl d.Y U".&.Y VI YLUYIU ..-. rimnlnt~+ tminino
a,
he successfuily completed the simu!ator check on May 12, 1992. He was assigned
5-3201 SIC duties on June 1,1992.
AKA Jetsiream 4100 training was conducted under contract with
Reflectone Training Center (RTCj, Sterling, Vire@nia, and included ground school
axd flight training for the J4101.
The captain's RTC 1-4101 instructolrs described him as an average
student. After completing 13.2 hours ir, f-4101 upgrade/transition trakhg, which
--.,-as his fist exposure to a "glass mockpit" type aircd, the captain failed his kitid

a
J-401 type rating check ride on September 30, 1993, because of difficulties with
instrument approaches, emergency procedures, and judgment. The Federal Aviation
Administration (FAA) examiner who adminisrewi fie faiied check ride reported
that the captain entered a pilot-induced oscillation while conducting an ILS
approach on standby inshuments during which the stick shah activated. He
observed that the =pain was 'husually nervaus" during the check ride.
The captain subsequently received 2 hours and 20 minutes of additional
flighi training on October 6, 1993. He succes!.fuBy passed a follow-up check ride
given by the same FAA exminer and received a type rating in the J-4100 OR
October 7, 1993. He was assigned 3-4131 PIC duties zs a reserve captain OR
October 26, 1993, after completing 21.2 hours and 1 I landings during initial
operating experieme (i@E). Pnor to his assi-merit as a 3-4101 reserve captain, he
had not served as 2 PIC in ai- carrier line operations.
At the request of the Safety Board, ACA provided a list of all fligbts
the captain undertook within the 90 day period before the acci-dent mad the time of
day the flights were made. This period included all fights he commanded as PIC of
the J-4100. Meteoroiogical information for thc conditions existing 2t the times, and
the locations of those flights was obtained to determine the nature of the weather the
captain lilkely encounrered at tile time.
The results indicate that during the previous W-day period, the captain
flew a totz! of 24 approaches to ten airports. Columbus, Ohio, was not one of these
airports. %e approach was flown when frozen precipitation was reported wit.?
iernpcrames below freezing. Xone of the 24 approaches were performed d e g
darkness, with frozen precipitation, in instrument meteorologicat conditions WC).
The captain had been offduQ ard at home dilring the 3 days prior to
the accident. His activities and meals, according to his spouse, were normal. His
rest was also normal: he rzceived apprcximatelq- 3 hours of sleep each night.

9
About 0200 on fanuary 7,1994, the captain was n&fiea by ACA crew
scheduling that his reserve sratus was being elevated to "A reserve status, which
required him to =port for duty within 90 minlltes of notification to facilitate
mtkipated crew requireaem. As a result, he took a taxi to Stamford and then flew
to Washington early on the morning of January 7, 1994, to NNI that r;eqUkement.
'FWe in the Washington, D.C., area during that day, he shared a hotel ma near
IAD witb the first officer of flight 291 and was later notified to report for the
accident flight The captain reported for the flight at 2021. No witnesses were
fmated &at could describe the captains activities while at the hotel, or indicate the
time that the captain went to k d on the night of famary 6, i934.
The fit officer of ACA f@ht 6163, the fli@ prior to the accident
flight, reported that after arriving at IAD, he secured the accident akraft and was
preparing tr, go 0% drrv when he encountered the crprain on the ramp and spoke
with him concekg the weatier. Tic first officer stated that the captain appeared
n o d in all respects. tht his demeanor vw dm ad professional, and that he was
'koncerned" about the weather.
A

a
10
cemficate with ratings and limitations for airplane single-engine and jnstmment
airpkme. ITIS total pilot time was 2,432.9 hours of which 110 hours were m
turboprop airplanes. At the time of the acci&ent, the fmt officer %it2 accrued a total
of 32.1 hours in the f-4101, of which 4 hours (iciucikg 1.5 hwxs on the accident
flighr) were We experience, and 28.1 hours were in flighr ixzkg, IOE, and
checking. He had a total of 16.3 hours night experience, and 11.5 hours & s t
experience t? the J-4101. The first oficer field a first class medical certifhte
issued on February 15, 1993 (second class medid cerltficate privileges after
August 1993), that cootzined no limitations.
Ihe first officer had ;lot: previously been employed as an air &.a
pilot. He had ken hired by ACA as a first officer 3 months befaz the accident He
had completed J-4101 secsz&k-comrmnd (SIC) pmd school and flight training
administered by RTC one m ~th prior to the accident.
The first officer's RTC instnrctors described him as an above average
student. He was given m oral examination by ACA on Ncvember 9, 1993, and a
SIC fiight check on Xsvemkr 22, 1993, and he passed the= both. Qn
December 9,1993, he completed 1 I hours and 10 landings of IOE traiiling in the
J4fOl and w2s zssigneci 3-4101 SIC duties as a reserve first officer.
?"ne first officer had been off dsty at his home in Bn;ok;yn, New York,
for 3 days prior to the accideni According to his spouse, he ate and rested nonnaily
&xig this period.

accident. He shared the hotel mom with the captain unti! he reported for the flight
Zi 2022 EST on january 7,1994.
A search of FAA records showed that &?e fjrst officer bad 20 accident
or violation history. A search of the first officer's FAA certification records showed
that on February 2,1990, he was unsuccessful in his first attempt to obtain his flight
inssructor's instrument certificate after failing the flight portion o€ the test. On
Febq 5, 1990, he passed his flight instructor's instrument flight test and was
hed a flight instructor's instrument certificate.
1.6 Aircraft Information
The airplane was manufittured on June 30,1993, by Jetstream Aimaft
Lhited, Prestwick, Scotland, as a mo6el J-4iC9, Serial Xum-ber 41016. 8n that
date, it was given an export certificate of airworthiness. British Aerospace, Inc.,
urn restructured in 1992. and, as of Tmuary 1, 1993, the portion of t5e company
*bt rnanEfactured turbopropeller aircraft was renamed Jetstream Aircraft Limited.
The airplane was acquireci by ACA in July 1993. It was certificated as
a J-4101 in the United SSRS as ;? msport-casegory airplme and was approved for
operat;on in icing conditions. day and night visld fli-ght rules (VFR), instnunent
flight rules tn;R), and reverse thrust taxi. Tfne airplane was configured to carry 29
passengers, two pilots, and one Right actendant. The airplane was equipped with an
autopilot, ground pmxiz-!ity war@ system (GPWS), CVR, and a digital flight data
recorder (FIX).

12
1. Use of the autopilot disconnect switch on either control yoke;
2. Pressing the Engage switch on the autopilot control panel;
3. Operating the electric trim switch on either yoke;
4. Operating the master autopilot switch on the glare shield;
5. Pressing the go-around button;
6. Activation of either stick shaker,
7. Pressing either FCS control switch;
8. Selecting reversion of any EFIS component.
The Jetstream series 4100 Manufacturer's Operating Manual, Volume 4
(&$OM 41, section 10, Autopilot, subsection F Captions (1) Ap DISC [autopilot
disconnect], states the fo!lowing:
0 Disconnection of the autopilot causes the AP DISC (red)
caption bo come on.
o An output to the Audio Warning System (AWS) causes a
horn to sound fol!.owing an autopilot disconnect. The horn
sounds for tu;v seconds irrespective of the eausc a? ~
disconnection.
o The AP DISC caption comes on for two seconds following a
pilot induced disconnection (e.g. stick shaker operation, GA
[go around] selection or deliberate disconnection) and is on
continuously followiig an autopilot sensed failure condition
or AHRS [attitude and heading riference system] failure. A
conti?uous warning is canceled by pushing the A/P OUT
switch.
On March 4,1934, Jetsrream Aircraft, Ltd., detemkd that portions of
&e preceding information exceTted from the kiOM 4 were incorrect Jetstream
issned a change on July 15, 1994, to correct the inaccuracies in the MOM 4
regarding the autopilot audio waning S ~SI~M
followkg a disconnect. According to
jetstream, &hp, audio warning system sounds for 2 seconds following an intentional
discomect by &e pilot (1, 2, or 5 h m the above autopilot disconnect list). Any
other type autopilot disconnect (3,4,6,?, 8) activates the audio and visual warnings
conti?uously until c anceled by the pilot.
~~

1.7 Meteorological Information
13
The 2200 National Weather Service (NWS) surface andysis chart for
January 7, 1994, depicted a low pressure area in southern Norii Carolina, central
Virginia and southex New Jersey. A cald front extended southwest into the Gal€ of
Mexico from the low pressure area in Virginia, while a warm front extended to the
east-northeast into the Atlantic. A trough of low pressure was located from
northwest Pennsylvania and extended northwest into Minnesota. A westerly flow
with high pressure was movlng into Ohio. The 0100 surface analysis for January 8,
1994, showed a low pressure area in approximately eastern Maryland with a cold
front extending to the southwest into ths Gulf of Mexico and a warm front extending
to the east into the Atlantic. A surface trough extended from the low pressure center
in Maryland northwest into Minnesota. figh pressure continued to build into Ohio.
me 23011 NWS weather depiction chart showed IFR5 and marginal visual flight
rules' (MVFRI6 conditions with snov throughout Ohio.
Aimm's rneteorologicai Laformation (WT) advisories for the
CMH area were issaed at 2045 and were valid until 03N on Jarmary 8, 1994, €or
occasional: LFK conditions in precipitation aiiijvi hg; G~cierzZ W %ii
between 8,000 and 20,ooO feet with isolated severe turbulence; and light to
moderate rime/mirred icing in clouds and precipitation between 2,000 and
19,ooO feet. There were no SIGMETs [si-gnificant meteomIogica1 information] in
effect for the time and location of the accident.
The &WS terminal wether forecast for Columbus, issued about 1945
and valid for the h e of the accident, was:
P
2000 to midnighk Ceiiing--800 feet c '%cast (visibility greater than
6 riles); winds 33G degrees at 8 hots; occasional cei!iig
1,200 feet overcast; visibi!ity 4 miles; light snow, fog.
Amendment 1, issued at 23@8, was as follows:

14
2308 to 0100: Ceiling--800 feet overcast; visibility 6 miles; fight
snow, fog; occasional ceiling 1,200 feet overcast; visibility 2 miles;
light sncw, fog.
Amendment 2, issued at 2354, was as follows:
2354 tc 0200: Ceilinp--500 feet overcasa; vi,sibi&y 2 miles; fight
snow; winds 3%3 degrees at 10 knots; occasional ceilig 900 feet
overcast; visibility 5 miles; light snow, fog.
The N WS area forecast (FA) for Ohio, issded at 2045 and valid until
January 8, 1994, at O900 was as folIows:
Northwest of a Cicinnati, Ohio (CVG) - Cleveland, Ohio (CLE)
line: Clouds 2,OOO to 3,000 feet brcken to scattered, 5,000 feet
overcast, fops 8,000 feet, with widely scattered visibiIities 3 to
5 miles in light snow showers.
Southeast of a CVG-CLE line: Clouds 2,000 feet overcast, layered
to 10,m feet, with visibiiities 3 IO 5 dies in iigiii snow, kg.
The surface weartier observations for Colilmbns are made by the NWS.
The NM-5 &ice is located about I 1/4 miles northwest of the approach end of
mway 28L. The foiIowing observatiorfi were recorded
Tirie-2250; type-record; ceiling-measured 1,100 feet overcast;
visibility-6 miles: weat!Ier-light snow, fog; temperature-23O F.;
dew poinr--22O F: wind-JIO degrees 6 knots; altimeter-
29.97 hches. ("his observation was transmitted to outside aviation
weather communica'ion circuits at 2251, transmitted on the
Autcmated Weather Information System (AWIS), and received in
the CM€i Air Traffic Control tower at 2249)?
Time--2305; typz-specid; ceiiing--measmd 800 feet overcast;
visibility6 miies; weather--light snow, fog; wind--290 degrees
'The time sWp on the AWS transmissions was slower than the actual time of the recorded
observation.

I994
15
6 hots; altimeter-29.96 inches. (This obsewatim was transmitted
on AWS at 2396 and was reeived in the CMH tower at 23C5).
Time--= 10; type--special; ceiiing--measured 800 feet overcast;
visibility-2 If2 miles; weather-light snow, fog; wind--300 degrees
6 knots; altimeter--29.% inches. (This observation was transmitted
on AWIS at 23 10 and was received in the tower at 23 10).
Tie--2328; type--special; ceiling--measured 700 feet overcast;
visibility2 1f2 miles; weather--1igkt snow grains, fog; wind-
230 degrees 5 knots; altimeter-29.96 inches. The snow grains
began at 2315. (This observation was transmitted on !-WS ?I 2329
and received in the tower at 2329).
Tine-234Q type-special; ceiling--measured 500 feet overcast;
visibility-:! 112 miles; weather--light snow grains, fog; wind-
290 degrees 10 knots; altimeter-29.96 inches. (The observation
was transmitted on AWIS at 2343 and received in the tower at
2333).
T~EE-Z~Q; Gp---~cord; ceiling-measured 500 feet overcast:
visibility-2 1/2 miles; weather-light snow grains, fog;
ternpcrat~re--23~F.; dew pint--22' F; wind--290 degrees I1 knots;
altimeter--29.96 inches. (The observation was tranmiiieC on
A\?TS at 2352 and received in the tower at 2352).
Freezing drizzle wjzs reported between OOO5 and 0033 on January 8,
The CMH h5VS wind gust recorder record showed wind speeds
varying from 4 to 6 knots between 2300 and 2310. The maximum wind speed
during this period was about 5 knots. Whd speeds increased to 14knots about
2337. The wind speed zt 2315 was about 5 knots; ai 2320 about 4 hots; and at
2325 about 4 hots. The wkd sensor is located about 115 feet north of the weather
service o%ce at a height of about 30 feel abve ground levei.
Upper air data for 19W, recorded at Daytor,, Ohio, abeu; 62 nauticzl
miles west of CMH, showed a temperature of about -5 degrees C at tke surface.
The temperatwe decreased to about -12 degrees C at around 8.000 feet. The wind

1994.
6 knots; altheter--29.95 inches. (This observatim was hnsmitted
on AWS at 2306 and was reeived in t!!~ CMH tower at 23%).
Tme--BlO; OF--special; ceiling--measured 800 feet overcast;
visibility2 In miles; weather--light snow, fog; wind--300 degrees
6 knots; altimeter--29.96 inches. (This observation was mnsrnitted
OF, AWIS at 2310 and was received in the tower at 2310).
Time--232S; tyqx--speci&, ceiling-measured 700 feet overcast;
visibility--2 1/2 miles; weather-light snow grains, fog; wind-
270 degrees 5 knots; altimeter--29.96 inches. The snow grains
began at 2315. observation was transmitted on F-WS nt 2329
and received in the tower at 2329).
T;Jne-2340; type--speciak ceiling--measured 500 feet overcast;
visibility-' ln miles; weather--light snow grains, fog: wind-
290degrees IO knots; altimeter-29.96 inches. Wie observation
was trammitted OR AWS at 2343 and received in the tower at
2343).
TL~c--~_~_~CI: ~ye--~cor& ceeiIing--measured 500 feet overcast;
vi~jibili~=-2 1/2 miles; weather-light snow grains, fog;
tempe~turg--23~F.; dew p0iqt--22~ F; wind--290 degrees 1 1 knots;
&i;neter--29.96 inches. (The observation was tran..&tted on
AWS at 2352 and received in the tower at 2352).
Freezing drizzle was reported between 0085 and 0033 on January S1
The CMH hWS wind gust recorder record showed wind speeds
varying from 4 to 6 knots between 2300 and 23110. The maximum wind speed
during this period was about 5 knots. Vv7i3d speeds increased to 14 knots about
2337. The wkd speed zt 2315 was about 5 knots; at 2320 about 4 knots; and at
2325 about 4 knots. The wkd sensor is located about 115 feet north of the weather
service office at a height of about 30 feet above ground level.
Upper air data for i9W, recoded at Cayton, Ohio, about 62 iiautical
miies west of Cbm, showed a temperatwe of about -5 degrees C at tie surface.
The temperature decreased to about -12 degrees C at araund 8,000 feet. Tie wkd

16
direction at the surface was northwesterly with a wind speed of about 3 knots. 'J%e
wind speeds increased to about 18 hots at S,W feet. The wind direction at
8,Gd feet was westerly. Moisture was evident from the surface up to aboul
5,OOO feet.
The weather idormation provided to the crew of N304UE included
departure and destination weather, winds aloft, alternate weather, en route weather,
and .WTs md SIGMETs. The information kcluded the 1950 CMH
observation of 900 feet overcast, visibility S miles; the CMH terminal forecast
indicating a forecast of 800 feet overcast, winds 330 degrees at 8 bo&, occasional
1,200 feer overcast, visibility 4 m i l e s in light snow and fog, which was valid fer &e
time of arrivak AEWETs for IFR, turbulence, and icing; and a pilot report zt 1911
indicating moderate rkrne icing at a FL (flight level) cf 4,OOO feet from CVG to
cm.
The a$& and fLst oS:cer of an America Eagle flight, a
SAAB/Fzk&ild SF-340. which landed at CMH about 2308, reported that during
their ILS approach io rsnwzy 28L. they encountered light-to-moderate rimehixed
ice. They we= provided a PEP biiot report] by air traffic ~0~001 of moderate
iciilg kiow 2,700fee~ and they entered clouds descending between 6,ocIo and
5,OOO feet. They cycled the deicing boots rhree times during the approach, and they
nozed, &ring a postnight kspection, about 1/2 to 3f4 inch of ice on the unprotected
surfaces of the airplane and no ice on the leading edges of the wings.
The pilot of a Hawker Siddeley HS-1000 landed on runway BL, about
2322, just befwe the accident airplae. He repor&ed that during descent, the
airplane entered clouds Setween 8,OOO and 7,OOO feet. He said that he encountered
light freezkg drizzle, light freezing rain, and ice fcg? and that the airplane
accrt?lu!&ed rim2 ice during the approach. The cirplane has 3 fillid anti-ice system.
He estimated a rate of accumulation of 114 inch for every 5 ininUtes of flight time.
Because of the ice, he added 10 knots to his asspeed. He slid that there were no
significant whd; during the approach. Tne pilot st&d that he broke out of the
clozas at 500 feet and that the ILS approach was noma! with nc warning flags. He
repor'td no ice acumiation on &e leagag edg: surfaces of the wing; however,
dmig pstflight inspectioa he noted 1/4 to 1/2 inches of ice on the nose of the
abpiane.
Two individrrals reported encounrering freezirg mist whiie driving near
t!!e aceideat si:e. One of them reported driving roughly parallel to the approach

17
ccurse near fie outer marker about 5 minutes befoie &e accident, a d the other was
driving to the accident site about 2340. Both of them j_r?scated diffi~ulty i~ &ing
because of the freezing mist sticking to their windshields.
1.8 Aids to Navigation
%at
Tiiere were no reported difficulties with the navigation aids used by the
at the time of the accident. A postaccident flight test Ori January 9, 1994, and
a ground check of the navigation aids, fsmd 30 malfunctions with the quipment.
1.9 Communications
At 230322. the flightcrew of United Airlines flight 660 that lmded at
CMH zdvised the Ixal controller of moderate ice between 2,500 and 3,000 feet.
The local controller *en advised the radar controller cf the pilot report so that this
inonnation co~id
be passed to subsequent arrivals. The radar coiltrofier wrote this
report on the back of a flightstrip and indicated the time o€ the receipt as 2336
AFhough he provided phis information, zs reqdired, to other aircraft that were
iandiiqg at CME, when the crew cf flight 291 made initial contact with him at
231Ck16, %e report was not passed to them.
At 23 1456, the ndar controller was relieved from the position, and he
briefed &e xelieving controller of the pilot report; hoaever, he neglected ts advise
her tha; tiis &formation haa Rot kea given to the €lightcrew of flignt. 291.
?here were no other ~ O Wair-to-ground R mmmunications dikufties.
The Pors Cohmbus InternationaI airport is 6 miles east of Columbus,
Ohio. at an eievarion sf 815 feet. Tbe airport has three runways: ~JI'O parallel
mways txiemec? iOR-28L and IOL-28R &mi 2 rhid runway, 5-23, isersects and
crosses rk.3 2% approach end. Rmway 28L is 10,250 feet in !en@ ami 150 feet in
wid%. Runway 28R is 6,900 fcef in length and 150 feet in width. Runway 5-23 is
4,383 feet in iength and 150 fee1 in wid&. Runway 28R is Gle prefemd rmway for
Usired Express due to the dose pmximiqf to its opera5ons.

approaches. Runway 28R has a localizer back course. Runway 5-23 does not have
an approved instmnent approach.
The airport has a fuli-time operational contrcf tower, ground mntrc:l,
clearance delivev, an6 ATIS. Air tnffic instrument appoach services to the airport
zre pruvided by Columbus approach control. The a‘Tort ab:, Z;as a UXICOM
frequency.
No reports or NOT’AMs [notice to aLmen], verbal or writtea, were
issued to the fli,ghtcrew regarding malf~ctiopls or improper equipment cn tht.
aitport. The pilot who landed just prior to the accident airplane reported that the
runway 28L runway lights were operating normally, althoagh they x :re not on ~e
highest intensity.
1.11 Flight Recorders
1.11.1 Cockpit Voice Recorder
The airplane was equipped with a CVR that recorded cockpit area
sounds. It was found about 2 feet fmm the separated empennage of the airplane and
was sent to the Safety Board‘s lahwatory in Washington, D.C., go? readout.
The majoriv of the CVR, a Universal solid-state digiial type recorder,
sustained severe fire damage, but impact damage was ger-erally confiied to the fro^:
side. Although all electronic components external to the crash case and thermsl
protection jacket were desiroyed by fm, no damage to .the memory module was
obsewed. The playback time of the recording was approximately 30 minuies and 1
second (30:Ol). The recording was of excellent quality enhanced by the use of the
intracocQit intercom system by the two crewmembers.
1.11.2 Flight Data Xecorder
“he postaccident fire was very intense, burning out of control for
nezly 1 hour, consdig a large portio3 of the airplane. As a result, the hits of
the solid state digital FDR’s ability to protect against fire were exceeded, and the
memory module suffered some thermal damage.
Tfie EDR, a Loral Fzirchild Model FlOW, was removed from the
airplane wreckage and was brought to the Safety Boards laboratory in Wxhington,

i9
\I;hle at the manufacturer's facility. the crash-survivable storzge unit
(CSS',:) was disassernbfed and kspented domn to the internal flash Iremxy storage
module. Repairs were made to she memory modufe, aIIowmg the recovery of &
record& data

approximately 16 feet t~ward the wreck>ge site.. A pair of similar ruts, that
exteaded about 3 feet, was four,d appximtely 85 feet fiom the fuselage and 2 feet
tu &e left of ae wreckage path centerline. "be lateral distance between the first
pair of NTS and %e second air of NS was about 20 feet. A third pair of ruts,
8 inches wide. started ;?bauut 74 feet from the fuseiage and was found in the m;ddie
of the two abve-rnentioned ruts. The= ruts continued for approximateiy 20 feet
and terminated n-ar *he was of the building. The aft tail cone structure was feud
where these i=ls s-med.

A-94-67
23

24
crewmembers abow the need to align the buckie insert to assure
easy release or" the safety belts.
A-9468
Issue a? Airworthiness DiPective to quire the removal and
replacement of &I safety belts mz~nf~cm-c.~ 5;. Pacific Scientific for
Part Number 1108435 buckles, with the 4.5' lift levers, md Part
Number 1108460 buckles with the 9@ lift levers, with belts having
buckles of a different design as expeditiously as possible, consistent
with the availability of replacement buckles.
A-94-69
Amend TSGC22f to incorporate procedures which wodd place
material representative of soft abdominal tissue between the test
appamus and the release hckle to emure that safety belts can be
released when subjected to loads specified in the TSO.
On June 5, 1994, the FAA responded to the Safety Board concerning
Safety Recornrendations A-94-67 ihrocgh -69. &I its reply to the FAA in a letter
dated September 1, 1994, the Safety Board classified these three recommendations
as follows:
a as
,*
@pen--
Unacceptable Response" due to the failure of the FAA to address
the need for operators to warn passengers of the possr~ility of inservice
buckles not operating pmpzrly. Safety recommendation A-
83-68 was c!assified as "Open--Acceptable Response" based on the
FAA's actions of issuing ai? Airworthiness Dictive to require the
removal and replacement of all safety belts manufactured by PacXc
Scientific with specific part numbers. Safety recommendation A-
94-69 was classified as "Open-Unacceptable Response," because
the actions takes by the FAA have not incorporated procedures
which would place material representative of soft abdomirtal tissue
between the test apparatus and the release buckle.
Safety recornmendanon A-Y.~-VI -* c- -.-- wa3 .,x,s.e. c ~ ~ o c ; ~
One of &e two passengers who was overcome by snoke and soot was
observed after the krplane c me to rest by the male passenger of the family of k e

1-16 Tests and Research
1.16.1 Flightpath Reconstruction
Port CoIumbus International Airport has an automated radar terminal
system (ARTS) operated by the FAA. An FAA-supplied magnetic tape containing
recorded data from ihe facility's computer was rezd out in the Safety Board's
laboratory. The radar ground track for the accident airplane was plotted s wig
3hut 7 minutes before the crash, when uze airplane was about 25 miles from the
runway, until loss of contzct at 232057.
The location of selected evems from the CVR and FDR are overlaid on
the radar data plots in fiws 2, 3, and 4. Each event location (dot) was
interpolated from the nearest radar data (circles). Radai altitudes were increased
75 feet to reflect actual altitide in feet msl.
Starting at the outer marker, the profile view of the approach is plotted
in iigures 5 ad 6. The local time, processed recorded radar altitude, FDR indicated
airspeed, and dialogue from the CVR are also shown on thf:se plots. The ILS
b Glideslope deviation recorded by the FDR is shown with the ~ angle syrnboi on the
plot. Because ody selected events are show, some FDR events or CVR sounds in
the cockpit are not shown on these plots.
The following is 2 brief description of the final minutes of flight data
recorder information:
At 231855, the "auproach capture mode" indicated a transition to
"capture" as the &S glideslope (G/S) value indicated less than
1/2 dot lww. The altitude was 2,988 feet and the airspeed was
I80 knots. The autopilot was "on." The data indicate passing the
cenkr of the outer marker 14 seccnds late: at an ahpeed of
178 hots as the altitude decreased through 2,784 feet, and the

26
NORTH RANGE - N.M.
Figm 2.--Seiected events from the CVR and &WR.

'i-
'i I
i
1 :
2875 0
27
& I i
2
-1.5 4.5 0.5 1.5 2s 3.5
NORTH RANGE - N.M.
Figure 3.-- Selected events from the CVR and FDR.

28
NORTH RANGE - M.M.
Figure 4.-- Selected events from the CVR and FDR.

UNITED EXPRESS FLW6291 DE§CENT PAT14
i i
n) 9
DISTANCE FROM RUMWAY - N.M.
Figure 5.--Profile view of approach.

500 1.000
'-I r
i
I
i
1
7.3 L
1.7 1
1
i
ALTITUDE - FT MSL
1.500 2.000
L' 1
OE
7

31
airpbne remained on the Iocalizer and G/S with the gear and flaps
in the up position.
At 2320102, engine torque vahes decreased from a previously
steady value of 25 percent tn t:&veen 6 and 11 percent, wMe the
propeller rpm values rerained steady at 97 percent. The airspeed
was 174 knots and Erm-easing. The flaps began to move from the
full-up position 4 seconds later, reaching 15' at 232025.
At 2320:42, the airplane started to descend below glideslope
I.? miles from the runway at an airspeed of 115 knots. CVR and/or
FDR data show that the Ian&ng gear were down, and wing flaps
were ai 15". Further, the altitude was approximately 637 feet above
runway elevation, and airspeed was 115 hots indicated airspeed
(KIAS) at 2320:42.
At 2320:45, the autopilot and yaw h-mr trmsitioned to "off" as
the airspeed decreased to 104 knots, and the radio altitude indicated
410 feet. The G/S data indicated that rple airplane was less than
1/2 dot !ow, as the AOA ad pitch attitude vzhes increased to
14.6' and 3.7', respectively.
At 232046.6, the airplane was about 2/3 of a dot below glideslope
when the stall warning system (stick shaker) activated. According
to company procedures, the minimum ILS approach speed at this
stage of the approach should have been 130 IUAS. The stick shaker
activated at 104.5 KIA3 and remained on for 2.9 seconds, until

32
2320A9.5: The stick shaker activated for the second time at
232051.0,101.5 KIAS, and 315 feet above the ground. FDR vane
angle-of-attack (AOA) values exceeded the stick pusher activation
threshoiri v^.+'O of a second after the stick shaker activated. The
-n
pun
-x-. CL-C P
JIIUWJ u1w q + t toques *beg&? tc, rise abme idle thrust at
232052.0, or 5.4 seconds after the stick shaker fmt aetkated. At
232054.9, FDR data show that the flap angle had started a steady
decrease th2t reached 0' by ground impact. Vane AOA values
rcpxt.-tq1y exceeded the stick shaker and stick pusher thresholds
during the final descent, until the airplane crashed at 232190. The
evidence indicates that pitch attitude and wing AOA were
bxeasing and decreasing in response to nose-up and nose-down
elevator deflections, respectively.
During the remaining 15 seconds of recorded data, the airplane
entered a series of pitch and roll oscillations, the power was
kc~eascd z& thz Szps we% xaised. The peak vertica! acceleration
recorded during this period was 1.52 "G," and lowest and highest
airspeeds were 99.4 and 124 knots, respectively. The peak torque
value w2s 84 percent recorded for the right engine 2 seconds before
the end of data. ne end of data was recorded at 2321:01, as the
pitch attitude indicated 22O nose up, ana m!l sttitude indicated 1.4O
right wing down
1.16.2 Fiight Tests
Under the supervision of the Safety Board, Jetstream Aircraft Cited
performed several flight tests to examine the high speed approacR techniques and
staII handling characteristics of the J-4100. The tests were carried out to:
1) Determine the stall warning speed appropriate to the
configurztion of the accident aircraft.
'Tis are reported to the nearest tenth of a second lrom 2319:32 until the end of the recording
at 2321:Oi.
fO§tra.?ght tine intergolation between data poi:ts was used to determine P;DR values to the nearest
tenth of a second.

33
2) Demomtrate recovery from autopilot disconnect at stall
waning during an ILS coupled approach.
The flight tests were conducted at the Jetstream facilities, Pmtwick
Airport, Ayshire, Scotland, in February 1994. The airplane med for the flight tests
was a current production J-4iW. The ody- nm~ai&-& i”tm L: &e &@ze we=
two video cameras fitted to record a general cockpit view and a close-up view of the
captaids electronic attitude director (or display) indicator (EADI). All relevant
parameters were recorded on the FDR to enable a direct comparison with the
accident airplane. The airplane was loaded similarly to the accident airplane.
Various stall approaches were flown to establish the effect of engine
anti-ice bleed and accelerated approach rate on stall warning and stall identification
speeds. Flight tests included: approacks flown that were similar to that of the
accident airplane; a denonstration of the free respowe of the airplane following
autopilot disconnect; and the effect of higher apprcach rates an8 lower flight idle
torques on stall speeds. The resuits were. then cornpared with the infomtion
svailable cr! the accident airplaae.
Recoveries from stick shaker were demonstrated in flight from both
conventional stall approaches and mupled ILS approaches. Recoveries from stick
pusher were also demonstrated in flight from conventional stali approaches, but ILS
approachss were simulated to maintain zafk aititcdes. ILS approaches were
simulated by placing the autopilot in vertical speed mode and allcwing airspeed to
decrease until stick shaker activated and disconnected the autopilot.
The airplane recovered without difficulty in each of the flight tests.
When recovery action (described i~ the I41 MOM) was taken immediately after
stick shaker, a further height loss of z.pproxi?ately 20 feet was experienced. The
height loss wben recovery action was not initiated until stick pusF : was
approximately 250 feet. Airplane response was also analyzed by flyiig to the
autopilot disconnect (stick shaker) and allowing the airplane to respond “hands
free” for approximately 5 seconds. Data from the flight tests was then compared
with that of the accidect aircraft which showed similar decreases in normal
acceleration, and nose-down pitch rates of approximately 2 to 3 degws per second.

34
Allowing for the effect of low flight idle (F/r) torques, high
deceleration rate, ice A0At1 on, and center of gravity corrections consistent with
the accident scenario, tle estimated stall warning speed would be 97.5 knots.
However, the stick shaker (stax waming) activated on the accident flight at 184.5
hots. According to the flight test report by Jetstream, the 7 knots had two
c~mpme~ts--&~~t 35 knots due to aerodphc lift degradation from ice, and
3.5 knots due to the high (2 to 3 knots per second) deceleration rate.
1.16.3 Systems Teardown and Examination
1.16.3.1 Ice and Rain Protection System
The J-4101 ice and rain protection system is designed to keep the
airpiane surfaces and the rain windshieids clear of ice and rain in all weather
conditions. The ice and rain protection system consists of:
Wing and stabilizer leading edge de-icing (boots)
Stall and AOA sensor x,J-icing (heat)
Elevator horn anti-icing (heat)
Totd air temperature (TATJ probe anti-icing (heat)
Engine air intakes anti-icing (heat)
Pitot and static port anti-icing (heat)
Windshield anti-icing and rain removal (heat)
Propeller anti-ichg (boots)
The wing, horizontal and vertical stabilizers are fitted with rubber
boots tkt inflate to break off accumulated ice. The boots operate from engine bleed
air pressure &at has been regulated to approximately 25 pounds per square inch
@si) and controlled by either a manual switch or autoritidiy via a the6 cirmit.
An ejector valve provides negative pressure to hold the boots along the leading
edges when not in use.
The stall AOA sensor is electrically heated and controlied by the left
and right air data switches. The case heaters, which ensure that $he stall AOA
sensors move freely in freezing condieions, are powered continuously, and the AOA.
sensors are activated concurrent with the air datz switches.
"See sections 1.16.3.1 and 1.16.3.2 for an expIanation of the AOA system.

The elevator horn anti-icing systems are electrically powered and
controlled by the engine air intake switches (ENGELEV --ICE). The syskm
receives inputs from the landing gear position and the engine-inlet anti-ice system.
The total air temperature (TAT) probe anti-ice system is electrically
powered and controlled by the left air data switch. The system receives input &TI
the landing gear position and waming system.
Thz air intake anti-ice systems on the left and right engines utilize a
combination of hot engine oil and bleed air from the engine. The left engine system
and right engie system are controlled by separate switches.
The three pitot tubes and two static ports are provided anti-ice
protection by electrical power controlled by the left and right air da& switches.
The main windshield anti-ice system is electrically powered by
separate inverters and cuntrolled by sqarate switches for &e left and right
windshields. The liquid accum&tion is removed from. ~e windshields using left
and @it windshield wipers. Contamination is mnoved by washing fiuid p ~~qed
tlmugh spray bars mounted on the wiper arm.
The propellers zrt: provided anti-ice protection by electrical power
controlled by switches md a timer. The timer mon~+-m and cycles the electrical
current to specific areas of the propeller anti-ice bocl.
The airplane is equipped with an Ice detection system. The system
gives a visual and audible waming of icing conditions. The system operates
continuousiy and includes a vibrating rod which is exposed to the airflow. As ice
ilcr;ul?iu1arcs,
--’-& Lb~-&u~&s
:-e 1 Ann I”.. -AA nnnnnt
w -....,. nrctain ” --_--_ the -_- ____. -nnm$nal frnpncy, md fie system
activates visual ard aural alarms in the cockpit.
The airplane ice protection systems were extensively damaged. The
wing deicing timer was found crushed and severely blnrned in the debris. No
information could be obtained &om the her. The pneumatic deich~!, and electricai
and bleed air anti-icing systems were examined €or failures. naalfunctions or
evidence of preiqact anomalies. The airplane’s ice contxol switches could not be
found. &?a pneumatic distribution valves were found. The left stabilizer pneumatic
deice boot was attached but was extensively heat damaged with iarge Ereas melteii.
Most of the right stabilizer pneumatic deice boor was consumed by fire. Only

36 ~
remmnts of melted rubber and small mas of boot material attached at the
peripheries of the stabilizer rmained. The stabilizer surfaces, where the boot had
been attached, Pxhibited scorching and extreme heat discoloration of the metal, wirlh
some melted alirminum areas. The vertical stabilizer pneumatic deice boot was
almost completely burned away with some boot material adhering to the top and
sides of the StabiIizer.
The windshieid wiper controls were not found. The windshield wiper
arms were found fire damaged and separated from their mounting shafts.
The elevator horn heated mats, the ice detector probe, and one of the
pitot tubes were examined. All of these parts were extremely fi and impact
damaged. No anornali.zs were noted.
1.16.3.2 Stall Warning Systems
A ground test procedure was performed on a sister airplane stall
warnir?g system that activated the left (No. 1) system. The left stall warning light
il!&%inated on the coming panels of both pilots, and the stick shaker activated on
the captain's control column. A similar rest was performed on the right (NO. 2) stall
warning system. When tested, the right-hand light on the coarnkg panels of both
pilots illuminated, and the stick shaker on the copilot's control column activated.
When both systems were simultaneously tested, the left and right lights
on the coa~tning panels of both pilots illuminated, both stick shakers activated, and
the stick pusher was enabled. The control column moved forward in response to the
stick pusher. A 65-pound force was required on the co~trol column to override the
stick pusher.
The autopilot was engaged in both the basic mode and in the coupled
mode. h i g either stall waning test, the autopilot became disconnected when the
stick shaker activated. Indications included an aural tone and the red -4P DISC
warning light.
During the air mode GC operation, the ground test features were
disabled and the sta!l warning identification functions were enabled, according to
the Jetstream 4100 MmrafacWr's Operating Manual, Volume 4. Tie ice mode of
the AOA system could only be enabled in the air when either the left or right
engine/eIevstor anti-ice system was activated and the airplane had exceeded

37
I45 hots. When the two conditions were met, a greea "ICING AOA." light
illuminated on the center instrument panel. When the indicator light was
illuminated, the system was in the ice mode, and the AQA at which stick shaker and
stick pusher occurred were reduced. To compensate for the reduced AQA inputs,
additional speed was added to the approach reference speeds for 15-degree and
254egree flap landing reference speeds: 12 knots was added for the icing AOA
15degree flap speeds, and 7 knots was added Sor the 25-degree fiap re." Azxnce
speeds, according to the ACA V-speed reference cards. The ice mode was disabled
if the engine/elevator anti-ice switches were tunred off and the STALL ICE MODE
PUSH TO CANCEL was depressed.
The examination of the burned wreckage failed tu disclose the
condition and/or operation of the stall warning system at the time of the acciden?.
1.16.3.3 Ground Proximity Warning System (GPWS)
The airplane was equipped with a GPWS that used the =-VI Ground
Proximity Warning Computer (GPWC) to provic'e alerts and warnings for
inadvertent flight into terrain. On Jeistream 4181 airplanes, the stick shaker inhibits
the aura! warning given by the GPWS, &iough the GPWS alert lamps are siot
inhibited and will i!luminate.
At radio altiixdes between 150 and 925 feet, a GPWS "glideslope"
callout will be heard when the airplane is o:n an ILS approach and descends
approximately 13 dots below the glideslope. At 232050.2, linearly interpolated
FDR data show that the radio altitude on the accident flight was approximately
339 feci as the ILS giidieslope deviation reached approximately 1.3 dots low.
Eowever, the stick shaker started less than 1 second later, which would inhibit all
GPWS cailouts. The glideslope callout was not heard on the C\X.
A "pull up" callout is generated by the GPWS to warn pilo'ts of high
dexent ates near terrain. FDR data from the accident flight show a radio aItimde
of approximately 208 feet and a barometric descent rare of approximately 2,500 feet
per minute when the sound of stick shaker ceased for approximately 0.6 second,
from 232053.7 until 232054.3. During this period, a "pull" callout from the GPWS
was heard, starting at 2320:54.8 on the CVR. According to the GPWS
manufacturer, after the warning envelope has been entered, the GPWC will
startdstop the voice callout rapidly with stick shaker deactivation/activation. There
is a 0.3-second delay for the "-VI GPWC (as installed on the accident airplane)

~~ .
. . .
38
to mognize that stick shaker has deactivated, and a 0.065-second deIay needed to
recognize that stick shaker has activated. The appropriate voice dout WS
st&stQp immediately aftex stick shaker off/on recognition takes place. Further,
GPWS voice callouts always start at the beginning of the statement, but f t ~
whenever the stick ;baker inhibit is pecognked.
cat off
According to tk CVR transcript, the "pull" callout started 0.3 seconds
after fie sound of stick shaker ceased, which is cmsisrent with the delay specified
by the GPWS manufacturer. The CVR indicated that the "pull" dout lasted
slightly less ehan 0.3 second. "&&ore, the ''pull" mllout by &e GPWS is
consistent with an abbreviated "pull wp" when the stick shaker activated for the third
fie at 232054.3 and 109 KIAS. The sound of stick shaker continued, and RO other
GPWS callouts were heard from 232054.3 until the enr?, of the CVR recording.
The GPWS on a sister airplane was ground tested. Upon activation,
the audible glideslope warning was heard followed by the audible "puli up," and all
lights illuminated in the GPW panel.
1
A subsequer:: test was conducted to include coupling the autopilot to
an ILS frequency on the ground at the airport, activating both stall wmdg system
tests and subsequently activating the GPWS test. The results included the
autopilot disconnecting; both stick shakers and the stick pusher activating; md the
GPVJS panel lights and stall identification Iights illuminating. lple audible warning
of "gii&slope" and "pul! up" were silent.
1.17 Additional Information
1.1%P Corporate History and Organization
On December 1 1989, ACA started as an east coast divisicn of
Westair Akbes, hc. ACA was located at Dulles Inremational Airport. In 1991,
the division was sold and began to operate as Atlantic Coast Airlines. The pwchase
included 22 BA-3291s and 12 EMB-120s. BA-3101s previously olpemted at hues
Airzort by Westair's East Coast division were exchanged for BA-YlOIs. The
certification process of the new company was completed by the Washington ]Flight
Standards District Office (FSDO) (EA-271, and ACA was ceMkated on
December 17, 1991, as a 14 C!?R Part 135 air carrier. iprn January 1, 1992,
operations began as United Exp~ss and served 35 cities.

39
The new President of ACA was the former President of the east coast
division of Wwtair Airlines, Inc. The positions of Senior Vice President of
Operations and Msintenance and the Director of Technical Services were added.
The Director of Operations was a former Vice President and Director of Operations
for Precision Airlines, Lnc. He also served as &e Assistant Director of Operatiom
a& Regional Flight Manager &lD) for Wc5tair ArSines, hc., prb: to the
incorporation of ACA. The Chief Pilot af ACA was a formcr pilot and Xegioml
Right Manager for Westair Airlines, hc.
In October 1992, ACA began preparation to qualify for a 14 CFR
Part 121 certificate to become effective April 1, 1993. The acquisition of 12
Depiavilland DHC-8 airplanes from Air Wisconsin required a Part 121 operatian.
The company submitted its request for certification under Part 121 to the
Washington FSDO in Jmuary 1993. me certification to operate as a conbined
Part 121/135 air came- was approved on April 1, 1993. The airline operated the
GHC-8 aiqJlanes to destinations previously served by Iir Wisconsin and to
destinations on the east coast.
During 1992, routes were extended into New England and Canada.
Service into Toionto was subsequently terminated by the corrrpany. In 1993, ACA
expanded as a United Express carrier into Florida and spcated six BA-3201 aircraft
in fhhat market. A pifot domicile was es!abIished at Orlando (MCO), Florida.
In January 1993, ACA notified the Washington FSDO of its intention
to place B.44101 (3.4101) airplanes on its certificate. A training department was
established specific to the BA-4101, and a Supervisor of Training position was
established for the BA-4101. Training materials were provided by British
Aerospace, Ltd., @Aej. The fmt aircraft arrived in May 1993, and deliveries
occurred at app.mximately 1 per month. A total of eight airplanes, incltiding the
accident akpiane, were delivered. At that time, the fleet consisted of 13 EME-l20s,
12 DHC-~S, 8 BA~LO~S, and 29 BA-3201s.
The company owned one Em-120 airplane; all other aircraft were
leased. At the time of the zccident, the compmy employed 312 cap’rains, 265 first
officers, 153 flight attendants, 126 licensed mechanics, ad 0 rmintenance
inspectors.

40
The bzses of operation were TAI) and MCO. A Rsgionat Flight
Manager supervised the MCO base and reported to the Chief SbL All aimaft
types v.ere opated out of L4D, and six BA-3201s were operated out of MCO.
The company had maintenance bases in Iilelboume (WB), Florida, for
BA-3201's, Lynchburg (LYH), Virginia, for all Jetstream eqspmnt, and Newbwgh
(Swr;), New York for DX-8s and EMB-I20s. SWF also served as a repair
facility. Line maintenance was performed for all aircraft at W. The route
structure was primarily east coast, north to south, serving abozn 50 cities.
According to the Director of Operations, the csinpany had a Supervisor
of TminAg for each type of airplane. Each Supervisor of Training had a flight
standards instructor for the particular 'p of equipment. The Director of
Operations served as the Director of Safety for both air and gruunc? operntiom. All
safety issues were brought befm the Hi@ Standards Advisory Boa&, which met
quarterly. The Flight Stmdards Ad-Jisory Board consisted of the Director of
Operations. Chief Pi!zt, Supervisor of Training, Hi@ Standards Instructors, and
company check airmen. Irreguhity reports aad a safety suggestion box were the
means by which safety-related issues could be communicated. At the time of the
accideat, there were no pilot reports xgarding safety issues on &s 3A-4101.
ACA system control was a 24-hour operation. System control
providsd a dual fimction: Flight dispatch for the Pzrt 121 Operations; and fI@t
folfowmg for the Part 135 operations. According to the ACA Flight Operations
Manual, Chapter VIE, page 1, system controii (flight dispatch) was operational
whenever sn ACA revenue fight was airborne and mzinbhed by at least one
licezxsed dispatcher. Cxvw pairings were monitored by a computer system used by
crew scheduling. Inexperienced crews, each wiLl fewer than loC! hours (not
including IOE) were considered "green on green." According to the manager of
system control, these pairings were sent ta the Director of Operations or the Chief
Piot for zppmval or disapproval. He indicated that there had been no pilot
complaints of "green'' crew pairiigs. Since the accident cmv was mt flykg xm&r
the "green on green" constraints, there was not a requirement for specific approval.
The Air b e Pilots Associatiof, (ALPAj represented the pilots of
ACA. The company and &e association had a Basic Eziploynent Agreement, dated
October 24, 1390. ori October 15, 1992, the antract was amended by Letter of
Agreement with regard to Sections 3, 4, 7, 8, 9, 10, 13, and 27. Accordt?g to
section 13, the bidding and $he filling of vacancies was based on the Pilots System

41
Seniority Zit subject to the basic agreement and the amendments. According to the
contract, wce a pilot was trained, he or she incmd a revenue service period (a
specific amount of time the crewnember must fly in that position for passenger
service). This period was based cn seat position arid airplane type. As a result,
more senior pilots were “locked “ into the captain and first officer seat positions on
the. 5-3201 and this provided &e ;jpportuni;y €or mare “junior” pilots to fill the
captain and first officer seats on the J-4101. Such was the case for the accident
crew.
1.17.2 EA-4180 Training
1.17.2.1 Ground Training
ACA Jetstream BA-4100 training was conducted under ccntract with
Reflectone TrzGng Center (RTC), Sterling, Virginia. The training division wa
formerly a division of BAe and was sold to RTC as part of a corporate restructuring.
RTC was still affiliated with BAe and provided all Jetstrea!! trainiag. This training
included ground school, simulator, and flight training for the BA-3201 md ground
school and flight training for the BA-4101. ACA used the center for both the
BA4101 and BA-3201 &a:hing. New hire pilots for ACA conmet with RTC for
their training and pay the costs associated wih the training directly to RTC.
According to RTC kstructors, ACA conducted aew hire, basic
kdoctximtion and Part 135 indoctrination. RTC maducted all aircraft ground and
flight training in a modular fom Tnitial pund training phases consisted of
64 hours. Forty-eigiat hours were aircraft ground training (Phase 1) and consisted of
airplane systems training. The infonmtion was taught from the ktstream 4130
Mmufacmrer’s Operating Mmuals {MOM) 1, 2, and 4. Aircraft gr0m.d training
(Phase 2), was also conducted by RTC. Gened operational subjects were covered
i~ Phase 2 from the MOMS, ACA Flight Operations ar-d Flight Standards Mmual,
and iasted 8 hours. fircraft ground training (Phase 3) consisted of four cockpit
pncedure W - g (CWj sessions, each lasting 2 hous The mockups consisted of
papr/photogaphs of the p=rn~ BA-4101 cockpit. The trainers were used for cockpit
orientation, profnes, flow and checklist practice.
The captain and first officer of flight 6291 attended a 1-hour c1as.r
during J-4101 ground zainhg &at addressed previous accidentsficidents, h:mm
factors,4mnsiderations9 and the National Aeronautics a?d Safety Adr&istration
(XASA) aviation safety repol-ting system. All human factors topics, including crew

42
Pn February 1994, ACA began prescsing a 1 .dq C W training mum
to its line pilots. The development ~f this course began prior to the accident. In
December €993, the fit officer participated for 1 day in a test class during &e
developmnt of the training course.
The syllabus RTC used was the BAe BA-410Q Training Mmual,
Chapter 4. The mama1 was used as a guide for ground, fight and CPT trai&g at
the center and in all phases of training: i~tial, tramition and recurrent.
Additional training conducted by RTC included 2 hours of airplane
emergency training for the pilors. Four hours of gmeeral emergency traiaing wzs
accomplished by ACA. The mining consisted of the interaction between pilots and
flight attendants while in simulated emergency situations.
1.17.2.2 Flight Training
ACA was the launch customer, and, at the time of the accident, it was
the SI$ operator of &e 3-4101 in the Unite0 States. The airplane WB newly
manafactwd, and a simulator has not been approved for pilo+ trainbg purposes.
The first simulator is scheduled for delivery to RTC in -Decenfmr 1994.
All training operations were condwted in the airplane fc . PICs and
SICS. The flight training consisted of 10 hours and a check ride for the I'TCs, and
12 hours md a check ride for the SICS. "he additional 2 how of training ic.r the
SICs provided additional tnsg for no&lying pilot duties. All 61ight instruction
was administer& by RTC inStrUCtQIS for ACA. Initial typ rating checkrides were
adiiistered by the FAA to the PICs, whereas RTC and ACA administered the
checkrides for the 5ICs. At the time of the accident, there were only two qualified
FAA J-4101 check pilots L? t!e United States. One was V~ed in the Washington
FSDO at IAD, znd t!\e other was based in Seattle, Washington. Upon wccessfkl
cornpietion of flight training, the pilots' miriing records were remrrmed to ACA, and
the pilors were. give? IOE by the airhe.
Accor~ling :e the ACA T~ir~ing Manual, SICs must receive 5 nom of
IOE in a pilot seat mder the supervision of a desigxated IOE check aimtant. 'Re

43
1,173 Altitude and Airspeed Awareness
A review af the manual prior to the change and removal of the section
on altitude callouts revealed the folh%ing:
1. The pilot not flying would cail out [approaching] 1w md
500' to any assigned altitude as a reminder to the pilot flying.
2. The pilot not flying wouid call out m y deviation of 108' from
any assigned altitude.
3. Altitude calls during instrument appuaches will be specified
i~ the approprizte Flight Standards Ma~ual.
An inspection of the night Standards Manual revealed no altitude
caiiout idonnation.

Atlantic Coast Airlines J41 PROFILES 341-19
?RA/NiNG P/&AUAL
Approach to Stall - Landing Configuration
REWSL?td n' 6
Date: March 75. 1933
Figure 7.-.J-4101 zpproach to stall -- landing configuration

45
A mah base inspection took place from Jmuq 29 to February 3,
1993. Ml deficiencies were comcted et that time, and 20 letters of corrective
actioa were sent to th8 company. Shiiaririy, a main base inspection was
accemplisbed on August 3 md 4,1993, with, similar conclusions m d outcomes.

46
Single i All Engines Operative !LS Approach
PF-’WS9’
%inirnurn Speed 1 40 KlAS
FLAPS 1-9.
LANDING CHECLC
(IF APPROACH NOT STASiUZEC BY 1000‘ MAA IMC OR 550’ HhA VM2 - GO AROUND)

ATLANTIC COAST AIRLINES, INC. was found to have deviated
from its approved or accepted procedrares in the areas of required
maintenance signof% on nomoutine work cards.
Compliance issues raised during inspection were discussed with
company personnel and the principal inspectoorx. Those issues that
could not be satisfactorily resolved, became fkdings in the body of
the report. In the case of findings where enforcement action is
anticipated, physical evidence and supporting documentation has
been provided to the Certifkate Holding District Office.
No fmdings pertained to operations training, crew qualifications, flight
control, flight operations, and operations records.
An initial Department of Defense POD) survey far ACA to enter 3to
$he DOD air Transportation Program was conducted between July 6 and 9, 1943.
ACA failed the sumey primarily because of maintenance issues. The airlime had
expanded from 30 to 59 airplanes, and there was Ett!e or no increase in ~laitem~~
supervisors for the increased number of flights. The airline was xevaluated on a
follow-up survey conducted between October 12 and 15, 1933, and all mas were
reported as average to above avmge.
Air Carrier Operations Bulletin (ACOB) 8-93-4 was issued by +&e FAA
on October 19, 1993, rega,rdiig "Flight into Pctential king Condiiions and the
Avoidance, Recognition, and Response to Tailplane Tce." The ACOB incsrpomted
several Safety Board safety ~ecsirmen&tioaa,s. (Set appemiix C). ??le PO4 for
ACA stated that he had received the ACOB and that he had a copy of it on file. He
said that fie believed the ACOB petained to J-3100 airplanes and he did not
issue the bulletin to the carrier since ACA did not operate this type of airplme.

A-94-70
48
Conduct an in-deRth review of its policies and procedures for the
processing of ACBBs, and devel.op a system ta ensure that the
safety infomation contained therein is acted on in a timely md
accurate m er. The system shsuSd include a process to vert@
that the actims contemplated by the ACBB are effectively
implemented.
A-94-7 1
Issue {mediate guidance to all POIs to verify that ~e intended
stkty-related actions contained in ACOB 8-93-4 have been
accomplished for air carriers under their jurisdiction.
A-94-72
Take the appropriate actions to verify that ACOBs issued in the
past few yem have ken implemented as intended.
&I general, the recomndations we= issued as the result of fidings
during this investigation and the Hibbhg investigation that revealed that PGI actions
speciflied in ACOB 8-93-4 had not been taken.

A-94-72 stated that the FAA will issue a notice directing its FSDO managers To
verify #at the actions contained in all ACOBs issued since January 1, 1992, have
ken accomplished. In its reply to the FAA in a letter dated August 3, 1994, the
Safety Board classified these three recommendations “Open--Acceptahle
Response.“
See appendix C for correspondence concerning thse safety
~ccammenciaiium.

2. B General
The flightcrew was properly certified and qualified in accordance with
applicable Fe&ral Aviatim Regulations and compmy requhments to conduct the
flight,
The airplane was properly ceMed and had been nnairniaint.6 iu
accordance with company and FAA requirements. There was no evidence of
preexisting discrepancies or preimpact mechanical failures of the structure, systems,
flight controls, or engines that contributed to the accident.
The forecast weather conditions pzovided to the flightcrew before
departure and during the flight were correct. The conditions called for XMC at FIE
time of the flight's arrival at CMH. About 6 minute.. prior to the accident, CMH
approach control advised the flight about revised ATIS information "Bravc." The
ceiling (800 feet) and visibility (2 1/2 miles) in light snow and fog contained in
Bravo was reduced fiem the ceiiig (1,100 feet) and visibility (6 miles) contained m
ATIS information "Mpha."
Before depar&ing UD, the flightcrew was given a PEEP indilicatjng
moderate rime icing at 4,090 feet from CY6 tc ClkM. The CVR transcript
indicates tht the flightcrew aIso received an icing report about 2300, approximately
21 minutes before the accident, fnm an airplane 25 to 38 miles ahead of them Tne
fiIighRtcrew of that airplane. indicated that they were expenacing moderate rime ice
up t~ 14,W feet and were in ?he clear at IS~OOG feet. About 2 minutes later, the
Eightcrew offlight 291 requested md received 15,ooO feet as a cruise altitude.

Si
the "boots" to dear it off. About 35 seconds after boot activation the first officer
stated "...little rime it never took nothing off this side here," to which the captain
agreed. The captain appropriately elected to conduct a "flaps 25 ice AOA on"
approach and Sanding. The flightcrew should not have experienced my signiPicant
diffkuities with tbe weather conditions during the approach and landing at CMH.
The evidence indicates that the captain of flight 291 followed company
procedures until the point at which he initiated the ES approach to runway 28L at
CME. HCWP.VP.T, he did not slow the airplane in sufficient time to be able to
configure the airplane in a timely manner. After reducing power to fight idle to
slow to approach speed, the pilots failed to monitor airspeed, and the captain failed
to add power as the airspeed approached 130 knots. The airspeed decreased
through the minimum of 130 knots for the appraach until the stick shaker activated
because the airplane was approaching stall sped. The captzin failed to execute a
proper stall recovery, and the airplane descended into the ground. Consequently,
the investigation focused on why the flightcrew failed to monitor the airspeed and
why the stall recovery procedure was not successFdi. Rightcrew training and
experience, company procedures, and FAA oversight of the operator were also
examined.
2.2 Flightcrew and Aircraft Performance
Although the ACA manuals did nct contain a def~tion of a stabilized
approach, the ACA mining manual did depict rn approved approach procedure
for the J-4101. The procedure depicts he airplane with engine torque at 30 percent
md airspeed at 180 knots before reaching the initial approach fix (UF) and after the
approach checklist is complete. It suggests a speed of 160 hQtS during the initial
procedure with no flaps. The procedure calls for the flaps to be extended to
9degrees when the airplane is established inbound on the localizer and within
3 miles of the final approach fk FAF) [distance measuring equipment @ME)
available], or when the fist indication of glideslope movement is called @ME not
available). The procedure depicts the airplane as configured with the landing gear
do-wn a d flaps set to 1 h Urpiee3 &a n LnFrrrn -IvIu tha .- find anrtrn~~h
yy.v *.. ti ..~d'po-~. int ~ ,- - ~ 1 ~ ~ 2 j - 1 3
At the FAFfFAP, the flaps should be lowered to 25 degrees with a n-&im&n speed
of 130 knots and engine torque at 30 percent. The procedure states ahat, at decision
13Tne procedure states that the flying pJot (captain in this case) should call for "gcar down, flaps
15, landing chcck," just prior to crossing the FAF.

52
altitude @A), Peduce sped to obtain the appropriate VEf at 50 feet above the ~
runway.
The accident fligki attained neither the configwtion nor the other
guidelines specified in the chat. About 10 miles from the airport, the airplane was
at a speed of 248 knots when the approach controller advised the flightcrew to slow
to 170 knots. At this point, the captain reduced the power to flight idie an& began to
slow the airplane. The airplane crossed SUMLE, the compass locator at the outer
marker &OM) and FpiF, at 178 knots with the airplane in a clean (flaps retracttd
r 1
and gear up) configuration. The hi& speed preventea &e crew
___.^ &--
ITJ ' 118 W'ffGlXS& ULflaps
to 9 degrees upon intercepting the glideslope and lowering the iariding gear at
the LOM. This was contrary to ACA procedures and ~~nstitt~kd an unstabilized
approach.
In addition, the power was re&iced to fight idle in a belated attempt to
lower the airspeed while descencliig on the glideslope. The law power setting
resulted in a rapid dzcelemtion, and without adexpate monitoring by the cxw, the
airspeed decreased hiow the 130 knots minimum required speed and below the
112 knots reference speed. Those speeds were based on the assumption that the
flaps wwld have been lowered to 25 degrees, rather than the 15 degrees of flaps
&at was actually achieved.
The autopilot was coupled for the approach, and the FDR dak show
proper localizer and glideslope tracking. However, the airplane was not eqipped
with an autothrottle system. Therefore, the pilot had to monitor airspeed md set
power accordiigly to maintain the proper airspeeds during the approach. The stick
shaker and stick pusher act as backup safety systems for pilots if they fzil to
properly monitor airspeed.
"he Safety Bsard believes that the captain xas aware of his airspeed,
hitially, because his call for "flaps 9 degrees" was comersurate with the maximum
airspeed of 170 knots. Similarly, the landing gear was placed down, md the call for
"flzps 15, !m"itg checks" was accomplished at appropriate airspeeds. However,
these calls and actions occurred 2 miles-or 40 seconds after crossing the LOM.
ACA does not have an approved high speed IlffsiVFR approach for the
H-4101. However, the training manual does contain a high speed ILS/VFR approach
published for the 5-3201. It states that the airplane sh~uld be slowed to 160 knots
approximately 3 nautical miles from the point where the descent is initiated, as

53
opposed to 130 hots for the standard ILS, appach. The CVR recorded the
captain stahg, "...keep her as dry as long as possible. We'll just ... briag her down
rea3 quick." T!x Safety Board believes that the captain probably wanted the
akpiane ts pass through the icing conditions rapidly, with the airplane in a clean
mr&gura!ion. The ckas configuration would allow minin;uum ice accretion while
passing thro@ &e icing layers. The captzin's flying experience dwkg the
predhg year was on the J-3201 as a fmt officer. As such, the Safety Board
belieyes that the captain would have been famibar with high speed approaches to
the FA!?. Although it was not an approved procedure on Lle 9-4101, it is possible
%z: %e -on+,sn
w-y .-- -v-*.-a
-- . -- !Q e mn6ifie.d 11-3201 procedure to mnetrate the icing
layers. The iwestigation determined that 34101 pilots do fly high speed
approaches for air traffic control considerations. However, this procedure is neither
published nor approved.
The ILS profde depicted in the flight manual 2lso coE'rzined a caution
that, "lf approach not stabilized by looo' ZPAA &eight abwe aimoft] MC or 500'
HAA VMC--Go around." Other than being estabIished on ^the Iocalizer and
glideslope, none of the depicted stabilized aprraach criteria regarding airspeed and
configuration were met when the airpiane passed through 1,800 feet W ic~ MC.
The captain did not begin to configurt! Lle airplane for landing until 48 seconds after
crossing the STT* -T. outer marker. At that time, the position of the airplane was
about 3 miles from the approach end of runway 28L. The final lmdig checklist
was not completed until the airplane was abaut 600 feet W, and the airplme was
not confipred for landing until that time,
The autopilot repatedly trimmed the airplane nose up to stay on the
glideslope, which, in conjunction with the low hst, caused the airspeed to
decay well below the minimurn approach speed of 130 knots. The CVR indicates
that less than 4 seconds after the captain stated, "and autopilot to go ... don't touch,"
the sound of the stick shaker began, followed by the tone for the autopilot
disconnecr. The airplane decelerated to 104 knots, which was 26 hots below the
mianinnurn approach airspeed specified by airline procedures, at which point the stick
shaker activated for 3.1 seconds. Immediately after the stick shaker warning, the
autopilot disconnected, and the airplane started to pitch down at approximately
3 degrees per second. Warning tones (presumably from the autcpilot disconnect)
started about (3.6 of a second after stick shaker. There was no dialogue heard on the
CVR until the stick shaker deactivated.

54
me evidence suggests && the csnptak was distracted by these events
and that he attempted to detem6ne what the fmi- oifisi k& done to cause the sfkk
shaker to activste andfor tl.rc autopilot to disconnect. During that short iritervai,
w5en the capkzii was 'nying to determine what hzd hapwned, the stick shaker was
silent. There was no indication on the CVR or FDR da& tkzt the aptain was aware
sf the extremely low airspeed and impending std! 'because he did not begin &e
proper sfadi recovery procedure. 'I%e cautaii ssked the first oficer, "what did you
rSo?" The f it officer respcnaed, "4 did& do nothing." Commensurate with the Erst
actuation of the stick pu5her, the power was partially applied to the engiws.
FIX cl&a indicate that the captain appiied nose-up eiev&toc -- w' :&x.,...+ lYi"UI
addiillg power. The airplane pitched up in mppcnse to the nose-up elevgar
ccmacci, but the airspeed was Too low to arrest ehe descent rate, an< "&e AOA
izxreased to the point tbt the stick pusher activated. The stick pusher quickly
moved the elevate: nose dow~p, which camed the airplms to pitch dmm, preventing
a stall. %wevt.r, FIX data. indicate that the cq&tin fought the stick pusher witb
large afl (nose-up elevator) control colurz inputs.
Engine power did not rise above id!e until 5 seconds after siick shaker
activation and .is seconds &er the stick posher activated. It %en increased only
about one-half as fast as would be expected from a full throttle a2piication. The
ezgine tomp reached 50 percent 10 seconds dter the f it stick shaker activation.
The captain then made a very serious error calilig for the flaps to be raised to zero
degrees. The stall speed for zero flaps is abwt 11 hots above the Caps 15 degrees
stall speed. Thus, the captain's action of raising the tiap and the failure to appIy
maximum power placed the eipiane within the aerod-ymmic stall region.
The initial response of the J-4101 flying pilot for missed approaches,
go arounds, and all approaches to stall in cruise, takeoff, or landing configuration is
nuximum power, flaps 9 degrees. In contrast to the approved procedure, about
1 second after stick pusher activation, the capkin called for "flaps up." There was
no further dialogue heard on the CVR until about In second a€ter the stick pusher
deactivated, when the captain stated "no no hold it," possibly in reference to the
previous flaps-up command. About 1 second after the stick pusher activated for the
second time, the capbin again stated "gimme flaps up."
The investigation revealed no procedure in either ohe 1-3201 or the
4-4101 in which stall recoveries or go-around procedures would requix a flaps-up
response. If the captaiq had reverted to previous J-3201 training for stall recovery

55
and misstated the command, the xespome would have been flaps 10 degrees (flaps
9 &,gees for the J-4101). Similarly, a flap setting of 10 degrees would have been
requkd for go-around procedures on the J-3201. "he delayed and insufMent
power appkz5sz revealed by the FDfP is inconsistent with the stall recovery
procedure.
The Safety Board corsidered the "flaps up" call by the captain in
connection with tailplane stall from icing. The vast majority of OAe captain's airline
experience was in the Jetstream 3101/3201 that previously had been involved in
tailplane stall accidents. Those accidents and their circumstances should have been
widely known by the pilots of these airplanes. The captain was obviously confused
by the stick shaker and autopilot wm-kags. It is also possible that the captain
believed the airplane was experiencing a tail stall. Such confusion and pcssible
misidentification of the problem couid have prevented the captain from
accomplishing the proper stall recovery procedure.
Xowever, the Safety Board discounted tailplane stall due to ice
accretion, and the captain's actions as being related to an attempt to recover from
tailplane stall, for several reasons. The 9-4101 horizontal stabilizer is designed with
negative camber on the upper surface to reduce the effects of ice accretion. In
addition, the boots have been extended farther back on the horizontal stabilizer to
ensure that any potential runback of ice can adequately be removed. Furtkr,
tailplane stall occurs as a result of a high speed with flaps extended rather than at
the lower speed at which the stick shaker actuates. Additionally, the proper
procedure to recover €rom tailplane stall in the 5-3100 and J-3200 was to add power
and remct the flaps to the mid-range pDsition. If the captain had perceived, in error,
a tailplane stall condition due to ickg, the reduction of the flap setting to a lower
angle would have beer, appropriate. However, the proper Bap callout should have
been "flaps 9 degrees," rather than the call €or "flaps up.*' Fiily, the aixplane's
pitch attitude time history obtained from FDR data was inconsistent with a tailplane
stall caused by ice. Consequently, the Safety Board concludes that the captain's
actions were not in response to recovering from a perceived *Alplane stall. The
Safety Board was unable to determine why the captain called for flaps up.
The Safety Board believes that the f3xt officer was confronted with an
increased workload for several reasons: the delay by the captain to configure the
airplane for landing; tasks associated with checklist completion; and interaction with
the captain. These activities supficiefitly distracted the inexperienced h t officer
and prevented hh from maintabkg awareness of the deteriorating pregress of the

With the stalI identification system ice AOA on, the stall w&g
system triggered at lower vane angles and correspondingly higher speeds. FDR
vane angle data indicate that the stick shaker and stick pusher both activated 3 times
during the accidcnt sequence. Pn each case, the evidence shows that the system

57
operated correctly, as designed. The following is from Jetstream’s flight test
report:14
The stall identification system is designed to provide warning of
approach to the stall and stall idenfllcation through stick shake and
stick pmh at pre-determined stall vane trigger angles. For an
aircraft with reduced lift m e
slope such as that of the xcident
aircrafi, these pre-determined trigger angles would result in lower
CL’S [lift coeffkients] and therefore higher stick shake and push
speeds. This would ensure that the pilot is warned and the stall is
identified and the required margins between the warning and the
$a!! are mzhah.n.t.,d.
The Szfety Board concurs with the conclusion reached in the flight test
report, which sun~xxizes the role of the stall protection system in this accident:
The stall protection system operate6 xxrectly and gave an earlier
(higher speed) warning of the impendhg stall througk %e operation
of the stick shaker appropriate to the loss of lift. The earik
waming was correct and mpropriate to an aircraft with an amount
I..
of accreted ice.
Flight tests showed that recovery was not difficprlt if power was added
promptly after stick shaker activation. The stall warning system provided a hely
waming to the flightcrew that the airplane was about to stall. In this case, the sFick
shaker activated 4.8 seconds before the stick pusher. This lead time was sufficient
to permit pilot intervention to prevent the q:d:l when the stick pusher activatcd for
the first time.
Prompt application of power and a swA1 aircraft-nose-down elevator
deflection would have resulted in a timely recovery from the low speed situation,
wiiiout activation of the stick pusher. However, no action was taken by the captain,
and FCR data show that the xcident airplane was only about 300 feet above the
ground when the stick pusher activated. A. successful recovery after stick pusher
activation at night and so close to the ground would have been diffcult, although
flight testing conducted at safe altitudes show that it was not impossible. From a
human perfonance standpoint, it would have been very difficult to maintain a nose-
14Jestream Aircraft Limited Flight Test Report, FIX 531/JiL141, May 20,1954.

down pitch attitude at night when the airplane was so close to the ground. In that
pegad, the captain's overriding of the stick pusher at t?at point, although ill-advised,
is understandable.
The Safety Board attempted eo determine the m e r in which two air
carrier pilots committed the fundamental errors that led to the accident. These
include the: 1) failure to monitor airspeed, 2) wisinterpretation of pronounced and
unambiguous cues of an imminent stall, and 3) improper stall recovery.
me evidence suggests that each crewmember possessed wipe
deficiencies that affected his performance during the flight. The Safety Board
believes that these deficiencies, alone or in combination, likely con&ibnted to the
errors noted. These include the captain's:
o documented history of poor execution of precision instmment
approaches,
o inexperience in nighttime, icing and restricted visibility
conditions in the J-4100,
0 inexperience with autopilot coupled approaches, and
0 inexperienx as a PIC.
The fmt officer, who was considered an above average pilot,
nevertheless, was:
o inexperienced as a first officer in schedule 14 CFB Part 135
operations, and
o inexperienced in the J-410.
The captain was concerned before departure about the weather
conditions en route and in the vicinity of Columbus. Evidence obsahed by the
Safety Board indicates that he had not flown either as PIC for ACA or as a
crewmember on the 3-4100 in the unique meteorological conditions present at the
time of the accident (darkness, low ceiling and visibility, fog, freezing temperatures,
and frozen precipitatioz). Further, according to the CVX during ff le execution of the
approach, the captain manifested apprehension about the performance of the first

officer. Perhaps it was these concerns that led to the tension iUwate0 by the minor
incidents recorded on the CVR. For example, at 2309:M; ’he captain accused the
first officer of giving him incorrect altimeter informatim. At 2309:44, the captain
vacillated on whether to direct the fmt officer tc obtain V,f with or without AOA
speeds for the approach. Pd 2312:29, the fist officer misinterpreted the purpose of
the captain’s readback of the assignee! altitude. Finally, seconds before impact, the
captain’s response to the stick sbker alert was to ask the firs: officer, “What did you
do?”
%though it is not unusual for pilots to become apprehensive in
ctralleeging flight conditions, air transport pilots are expected to execute their
;;iioting skills and to display jud-pent independent of whetcver stress or tension
they ma.? be experiencing. By contast. the report of the cxamAer who acministered
and failed the captain on his initial type rating ride on the 1-4881, for inadequately
executing an Its approach, hdicates that th?s captaiz’s perfbmance deteriorated
when he became nervous. The Eature of his piloting emrs on ths night of the
accident, especiaiiy his failure to monitor airspeed, is consistent with &dings of
humaa factors research indicaciog that excessive tension can predictably degrade the
ability of human moritors to obtain and integrate idomation from multiple sources,
a phenomenon referred to as ”attenticma1 narr~wing.”~~
The Safety Board examined the display of airspeed within the
airplane’s electronic attitude director indicator, a cathode ray tube or CRT, to
determine if the m e r in which the information was presented codd Pave
hampered the ability of either pilot to perceive and integrate the critical information.
Airspeed data on the 1-4100 is presented digitally on a vertical moving display, wish
the airplane’s indicated airspeed centered within the display. As the airspeed
increases or decreases, the displayed airspeeds move up or dom correspondingly..
(See fi&ure 9).
’Ibis format is similar to that of electronic displays of newer generation
“glass cockpit” aircm? that have Seen introduced into service witkin the last 8
years. Pilot acceptance of the displays has generally been favorable, and, more
%+!Ortant, they have not been suggested as contriburory to accidents. Moreover,
their presentation format across aircraft types has generally been consistent with
hman factors priiciples of presenting visual information. For example, in the
15Hanccck FA., and Warm, J.S. (1989) A dynamic model of stress and sustained attention.
Human Factors, 26,519 - 537.

low ainpeed
awarenzss indicator color]
60
Figure 9.--EIectrsnic attitude director indicator.

61
5-4100, the moving vertica! display presents trend information, and, as the airspeed
approaches the stail speed, the color of the dispiay changes to red, the common
color of waning. Therefore, because the airspzed display on the 5-4180 was
consistent with these principles, the Safety Board does not consider their format or
mode of presentation tc be a factor in this accident.
Ne.verlheless, the circumstances of tlnis flight shocid not have been
especiaEy anxiety provoking. Nighttime fligh't in icing IMC conditions, although not
routine, were not beyond the ability expected of air transport pilots. However, in
this accident, the evidence indicates that the captain's own fail% to stabilize the
airplane on fmal approach, in accordance with the flight profile suggest& in the
ACA MOM, likely created the circumstances tkat exacerbated the anxiety he was
experiencing. Thus, when the airplane was estab!ished on the localker, and its
airspeed and configmation were stili not stabilized, the evidence indicates that the
captain was overwhelmed by a need to perhm certain actions simuhneousiy.
These included the need to:
C slow t!e airplane down and establish the proper airspeed,
0 maictain a precise flightpath in the restricted visual conditions,
3 maintain a vigilance for ice accretion, and
o closely cbserve the first officer as wel! as manage his acticns.
The evidence indicates that the captain was unable io perform these
actions when required. Further, his use of the autopilot to help with one of the
required xctions, rnairm*&?iEg a precise flightpata?, suggests an unf&Earity with its
capabilitics. This suggestion is supported by the capraink reported consistent use of
the autopilot when flying instrument approach profiles. Given the deficiencies in his
execution of instmment approaches, as documented in his check rides with ACA,
the captain appears to have demonstrated. especially when nervous, a weak
instrument scan in high perfsrmance aircraft during restricted visual conditions.
A pilot wi?h a poor scan could rely on the autopiict to fly a precision
hstrument approach, with the knowledge that the sqstem should reliably and
accuztely execute both the glideslope md localizer flightpaths. However, use of
the z~topflot~ without compensating efforts to thoroughly monitor necessary airplane
instmnent.5, could exacerbat? a possible poor instrument scan, since no effort is

61
5-4100, the moving vertica! display presents trend information, and, as the airspeed
approaches the stail speed, the color of the dispiay changes to red, the common
color of waning. Therefore, because the airspzed display on the 5-4180 was
consistent with these principles, the Safety Board does not consider their format or
mode of presentation tc be a factor in this accident.
Ne.verlheless, the circumstances of tlnis flight shocid not have been
especiaEy anxiety provoking. Nighttime flight in icing IMC conditions, although not
routine, were not beyond the ability expected of air transport pilots. However, in
this accident, the evidence indicates that the captain's own fail% to stabilize the
airplane on fmal approach, in accordance with the flight profile suggest& in the
ACA MOM, likely created the circumstances tkat exacerbated the anxiety he was
experiencing. Thus, when the airplane was estab!ished on the localker, and its
airspeed and configmation were stili not stabilized, the evidence indicates that the
captain was overwhelmed by a need to perhm certain actions simuhneousiy.
These included the need to:
C slow t!e airplame down and establish the proper airspeed,
0 maictain a precise Sightpath in the restricted visual conditions,
3 maintain a vigilance for ice accretion, and
o closely cbserve the first officer as wel! as manage his acticns.
The evidence indicates that the captain was unable io perform these
actions when required. Further, his use of the autopilot to help with one of the
required xctions, rnairm*&?iEg a precise flightpatal, suggests an unf&Earity with its
capabilitics. This suggestion is supported by the capraink reported consistent use of
the autopilot when flying instrument approach profiles. Given the deficiencies in his
execution of instxvment approaches, as documented in his check rides with ACA,
the captain appears to have demonstrated. especially when nervous, a weak
instrument scan in high perfsrmance aircraft during restricted visual conditions.
A pilot wi?h a poor scan could rely on the autopiict to fly a precision
hstrument approach, with the knowledge that the sqstem should reliably mild
accuztely execute both the glideslope md localizer flightpaths. However, use of
the z~topflot~ without compensating efforts to thoroughly monitor necessary airplane
instfarnents, could exacerbat? a possible poor instrument scan, since no effort is

62
needed to relate the airplane’s flightpath to corrections to the control surfaces
necessary to maintain the proper flightpath. The evidence sliggests that this
captain’s use of the autopilot exacerbated a tendency to a deficient instrument scan.
That is, his performance on the accident flight, and repcpts of his use of autopiloS
during the execulion of insmment approaches, indicated that he relied on the
autopilot to maintaitl a stabilized flightpath without concomitant monitoring of the
critical airplane parameter of airspeed, a pmeter not contloiled by the autopilot.
It is possible that both pilots, given their reldive inexperience in the J-
4100, were not suffciently experienced in the digid format in wfiich airspeed was
przsented to provide them with the necessary ability to mtieipzte critical trends in
the airspeed. Bxause both had considerabiy more experience piloting aircraft with
traditional electromechanical instrumentation than with aircraft with “glass cockpit”
type digital presentation ~f flight idomation, the Safe? Board could not: d e out
their relathe inexpienee with electronic flight instramentation as a potentially
adverse influenct on their performaace on the night of the accident
When the stick shaker alerted, the CVR established that neither
crewmember recognized that the aimiane was about to stall. They failed te focus os
the airspeed, after the stick shaker alerted, and neither commented on nor displayed
a recognition of the airplane’s precarious airspeed simation. The captain‘s
vacillating cdls for flap retraction fiihher illustrate his unawamess of the ahpeed
and the meaning of the stick shzker. Although it is GEicuit io explain how an zir
kx~sport pilot. could not respond appropriately to a stick shaker, it is apparent that at
that point in the flight, bath the captain and the first officer were unaware of
fundaxental parameters and unable to anticipate the airplane’s flightpath. Thus,
they were “behind the airpiwx? and unable to plm and control !he airplane’s
flightpath a ~ to d respond appropriately to tfe stick shaker.
1.n summary, the evidence suggests that the combinatior. of
inexperienced first officer, nighttime, restricted visibility LI icing conditions,
inexperience on the 3-4301 and the use of its autopilot, and inexperience as a FIX@,
contributed to the captain’s failure to monitor the airspeed, once the airp!me was
estabiished on the approach. The failure was most likely caused by atter$iondt
narrowing as a rextion to the stress the c;tp+ain experienced while flying the
approach. As a result, when the stick shaker alerted, neithel- cmvmemkr
recognized tht the airpime was about to stall, and neither appreciated the need for
the implementation of prompt and appropriate stall recovery techniques.

63
2.3 Pilot Training and Exgeniense
The Safety Board reviewed the training received by both L’le cap+&n
and the first officer. The ground training and flight training reqsirements met or
exceeded the minimum requirements as set forth in Fedemi Aviation Regulations.
Interviews wkkk ACA pilots and FAA personnel revealed that RTC had
wellqualied flight and growid instructors. Since the airplane was new to the
United States, rhe training faciliiy was constantly modifying and making
improvements to the training program. RTC had a designated POI for the facility
&at provided oversight for the FAA. He had not reported any deficiencies in the
tnining program for the J-410 1.
At the time of the accident, there was no J-4101 simulator available for
iraking anywhere in the world. The first simulator is schedoled for operation in
December 1994. As such, all training, at the time of the accident, was accomplished
in the airplane. Pilots interviewed stated that the flight training was excellent. The
company check airmen interviewed stated that the transition during IQE was easier,
since pilots had actually flown the airplane. None of the pilots interviewed
indicated that abnormal or emergency procedures that were simulated in the airpiane
resulted in a poor learning situation or lack of knowledge transfer. Nevertheless, the
§afety Board believes that the lack of a simulator, specifically designed for the J-
4101 airplane, hits a pilot’s training and subsequent ability to perform certain
procedures that can only be safely practiced in a simulator. For example, stick
shaker activation dming instrument approaches would not be a safe practice during
mining flights in the actual airpime.
Auto?ilct-coupled approaches were listed as part of the flight mining
requirements for some of ACA’s airplanes (DHC-8s and Em-I OS), and the ACA
trailing manual covered the 3-4101 autopilot as a subject in ground training.
Fbwever, the investigatios reveaIed that autopilot-coupled approaches were not
fisted as a specific &aining event in the ACA J-4101 flight training manual or on the
flighr eiiaiuafion fom. For standardizatiou, a revised flight training evaluation form
was printed to include all the airplane types ACA operated. Autopilot-coupled
approaches were an item prinred on the Form. The form was printed with a revision
date of 3a!y i 5, 1393. Although training was accomplished by both CI zwnernbers
after that date, an earlier form was esed thzt did not list autopilot-coupled
apprDaches.

63
2.3 Pilot Training and Exgeniense
The Safety Board reviewed the training received by b@th the cap+&n
and the first officer. The ground training and flight training reqsirements met or
exceeded the minimum requirements as set forth in Fedemi Aviation Regulations.
Interviews wl::n. ACA pilots and FAA personnel revealed that RTC had
wellqualied flight and growid instructors. Since the airplane was new to the
United States, rhe training faciliiy was constantly modifying and making
improvements to the training program. RTC bad a designated POI for the facility
&at provided oversight for the FAA. He had not reported any deficiencies in the
tnining program for the J-410 1.
At the time of the accident, there was no J-4101 simulator available for
iraking anywhere in the world. The first simulator is schedoled for operation in
December 1994. As such, all training, at the time of the accident, was accomplished
in the airplane. Pilots interviewed stated that the flight training was excellent. The
company check airmen interviewed stated that the transition during IQE was easier,
since pilots had actually flown the airplane. None of the pilots interviewed
indicated that abnormal or emergency procedures that were simulated in the airpiane
resulted in a poor learning situation or lack of knowledge transfer. Nevertheless, the
§2fety Board believes that the lack of a simulator, specifically designed for the J-
4101 airplane, hits a pilot's training and subsequent ability to perform certain
procedures that can only be safely practiced in a simulator. For example, stick
shaker activation during instrument approaches would not be a safe practice during
mining flights in the actual airpime.
Auto?ilct-coupled approaches were listed as part of the flight mining
requirements for some of ACA's airplanes (DHC-8s and Em-I OS), and the ACA
trailing manual covered the 3-4101 autopilot as a subject in ground training.
Fbwever, the investigatios reveaIed that autopilot-coupled approaches were not
fisted as a specific &aining event in the ACA J-4101 flight training manual or on the
flighr eiiaiuafion form. For standardizatiou, a revised flight training evaluation form
was printed to include all the airplane types ACA operated. Autopilot-coupled
approaches were an item prinred on the form. The form was printed with a revision
date of 3a!y i 5, 1393. Although training was accomplished by both CI zwnenbers
after that date, an earlier form was csed thzt did not list autopilot-coupled
appmaches.

64
The former POI, a J-4101 type-rated FAA inspector who gave the
captain his type rating, stated that he preferred EO see a candidate demonstrate
ability in using the autopilot during checkrides, since many of the pilots had no
autopiiot experience prior to the J-4101.
During the qualification check ride, the former POI required the captain
to demonstme satisfactory autopilot knowledge while flying a coupled approach.
The general consensus of Reflectone instructors and the FAA was that many pilots
hired by ACA had aviation backgrounds that did not inciude the use of an autopilot.
Bscausr: of this, it was necessary to train and check the use of the autopilot.
The Safety Board believes that although adequate autopilot training
was accomplished by the KTC and that it was aciequately addressed by the FAA
during checkrides, the incorporation of an autopilat-coupled approach training item
in the ACA flight training manual and the Reflectone syllabus wouid preclude the
passiloiiity of coupied approaches beirig o\-&a&&. FxE!x~,
tc: inrl~de mtnni!nt-
coupled approaches as a~ item on the ACA pilot proficiency check form would
ensure that pilot knowledge and use of the autopilot during coupled approaches was
reviewed.
The Safety Boarc! believes that experience gcined as a first officer with
an airline, prior to upgrading to captain, is important. Contract training instructors
may possess considerable air carrier line operating experience: however, the airline
should be the final determining factor in pilot qualification for line flying. Li this
case, the captain of flight 291 went directly from first officer in a Itis sophisticated
airplane (J-3101132Olj to captain of the 5-4101.
Although Jetstream manufactures both the 5-3201 ar,d the 3-4101
airplanes, the differences between the two airplanes are significant. Th;: 5-4101 is a
newer and more complex airplane. The addition of an autopilot and modern
instmmentation (glass cockpit displays) are two of the major differences between
the airplanes. A pilot transitioning from the J-3201 to the J-4101 could not apply
previous system knowledge (in cockpit layout or design) learned in the 5-3201 to L??
3-4101. There are very few similarities in ai.rplane systems. Funher, the
instrumentation in the J-320: is analog, whereas the Instwnentatlon in the 5-4101 is
an electronic flight instrument system (EFIS). The new glass cockpit design
requires the pilot to learn a new concept of instrumentation. The investigation
revealed no pilot comments regarding difficulty in flying or intqreting the EFT$
sysrern installed on the 5-4101 airplane, and EO comments we;e received to imply
f

that the captain or the first officer involved in the accident were deficient in
instrument skills usin.g the EFTS system. However, both pilots were inexperienced
in the. new airplane aad failed to scan the insmments properly during the high
workioad of the accident flight.
Pilot evaluations of Reflectcne's flight training were very favorable.
The contract instructors were considered experienced. Training during shll
procedures varied somewhat among instructors. Some of them allowed the student
to proceed past the stick shaker to t3e stick pusher, whereas other kstmctors
demonstrated to the stick shaker oniy. In either case: the student had to demonstrate
knowledge and competence, both to ihe instructor and the FAA, regarding stall
recovery. The stall training had to be conducted at a safe altitude and not during
actttal instrument approach conditions as could have been demonstrated in a
simulator.
Because the captain was newly upgraded to PIC on J-4101 airplanes
and the first officer had been hired 8 months prior tc the accident md completed
5-4101 pilot training within 2 months of the zcciknt, the Safety Board believes that
it is unlikely that either crewmember was adequately seasoned in his respective role.
The captain had been employed by ACA as a first officer for more than I year Gn
the 1-3201, prior to his upgrade to captain on the J-4101. Since the upgrade, he had
served just over 2 months as an ACA 3-4101 captain, The first oficer had
completed new hire ground and flight training abut 1 month prior to the accident
and had flown only one round tf? as a first officer in ACA passenger operations.
The captain and first officer were fiends; however, this was their first flight
together.
The investigation determined that the captain expressed concern, prior
to departure, about the en roufe weather, turbulence, and related icing conditions in
the vicinity of the airport at Columbus. The CVR indicated tht the captain
adequateiy addrcssed these conditions during the course of the flight. AI' interview
with another ACA copilot, who had flown with the captain for 15 aqs in
December 1993, indicated that the captain frapently iikeci to couple the airplant: to
the autopilot, on approach, rather than to fly the airplane maually. A review of the
captain's records indicated that the two failed checkrides (SIC on the 5-3201 and
PIC on the $-4101) were, in part, due to uosatisfactory performance on approaches.
On subsequent rechecks, he demonstrated satisfactory proficiency after retraining.
The Safety Board believes that t;?e caprain was inexperienced and iacked confidence
in his ability KO fly &c J-4191. but that he was aware of his weaknesses. As a result,

he may have relied on the autopilot to supplement his flying abilities and erLhaxe
the apprcach stability of the airplane i n less rhm optimurn weaxher conditions.
The Safet;v Board acknowledges the valw of an autopilot to FL&IC~:
pilot wnrkioad during instrument approaches and encowages its use. However, &E
Safety Board is concerned that some pilots might accept autopiior performanr.:. as
infallible and become complacent in their monitoring function. ?he Safety BG&
belleves that training programs must stress the need for pilots to stay &,rt md
remain in the loop durhg coupled approaches.
The events of this accident reelect a !&a1 breakdown 13 crew
codination, an essential element of conducting successful instrument approaches.
CRM training is not currently required ~nder 14 CFR 135; nonetheless. ACA did
include a I-hour class during its J-4101 ground school that inc!.l-fed previous
accidentjincidents, hcman factwkonsldemtions, and the NASA aviation safety
reporting system. The training did not provide for interaction of the crewmemkrs
or feedback and continued reinforcement regarding their peSomance, as described
in Advisory Circular (AC) 120-5 I A Crew Resomce Management Training.
A.tlmtic Coast Airlines began operations on 9af:uary 1, 1992, with a
rnm2gement spructtlre experienced in airline operations. That experiencz, according
to the FAA POI, enabled the company to avoid many problems that new entrants
had in st~t-~p airline operations. A rapid expansion occurred within &e company
during the foliowing 18 months.

67
ACA had a seniority system for the pilots; however, pilots that had
'ban trained in a particular cockpit position or type of airplane incurred the "iock"
or "freeze" due to the revenm service perid. As a result, junior piiots were able to
fill a position on a higher paying airpiane model. Such was the case of the accident
captain. He was trained as a first time captain to fly on the 5-4101 and, in effect,
j':nped ahead of other pilots, out of seniority, on a aew and more desirable piece of
p.;uipment. The newly hired first officer was also placed on the more sophisticated
piece of equipment.
Although a pilot seniority system does not guarantee that the most
qualified airmen are promoted fit, the Safety Board believes that ser'iority does
providp ari indicator of seasoning and experience in airline operations. The system
in place at ACA precluded the orderly progression that would have enabled the new
captain to gain experience as a captain on a familiar airplane before helshe
progressed to a new and more advanced airplane.
Al?houg'n the company met or exceeded the ground and fight training
requirements and regulatims, the operational oversight and monitodg of the pilots
by company managers appeared to have been reduced. ?he lack cf adequate
supervision md guidance may have led flightmws to develop poor gig?&
procedures and habits. An example was the procedure of flying high speed
approzches to assist air traffic control. The nunstandardization of operations
between airplanes was recognized by management and was being addressed by the
cornpzny through the deveIopment of a Bight standards manual. At the time of the
accident, the manu& had not been zpproved by thz FAA. While the captain had
more fiight experimce than the first officer, he had been recently promoted Erom a
first officer on a 3-3102 to a captain of a 5-4101 on a scheduled air carrier. If
standardization of qproach procedures between airplanes had k en established, the
captain might have been better prepared to carry out proper approach procedures,
and the first officer might hzve been more knowledgeable and trained for the event.
The company correctly applied the "green on green" pairing
:xxqtitirements of pilots not flying together with less than 100 hours. However, the
combination of a new capain with a previo?~ history of demonstrated yroblems
d~ri~:; checkrides. scheduled with a new first officer who had not flown for 18 days,
prsvided a degraded Rying performance environment that proved to be inadequate
usder the existins operational conditions.

68
As a result of the Safety Board's Safety Recommendation A-90-iO7,
originating from the investigation of USAk flight 5050, i6 on July 22, 1991, the FAA
issued ACOB 8-88-1. This ACOS revised existing guidance concerning &e pairing
of crewmernbers by incorporating a joint govemmenthnduswy task force's
flightcrew perfomaxe committee recommendations. The c:rJr.mittec's
recommendations involve three basic program elements: consolidation of skills,
operating reslr;actions, and pairkg restrictions.
The Safety Board investigation indicated thae &e FAA surveilance of
ACA was conducted in accordance with flight standards directives. After the initial
certification to operate was issued in late 1091, the TCI was infcmed of- the iwnt
by the carrier to expand operalions under Part 12 I. A schedule of events to certify
the carrier was conducted and completed in April 1993. Additiofially, &e carrier
submitred is request to place the Jetstream 4100 on ita certificate. Again, the
certification process for inclusion of a new airplane was accoTnplished at ACA. The
oversight by &e FAA during initial certification and during ?ne recertification for the
additional company operations was adequate. The Safety Board believes that &e
FAA's role in approving the carrier's operating certificate for the Jetstream 4100 was
proper and did not contribute to the accident.
2.6 Corrective Actions
As a result of the Safety Board's ir'vestigafon of the GP Express
accident in Anniston, Alabama, on Apri! 12, 1993. the following recnmmendation to
the FAA was issued: I7
'4-93-36
Require that scheduled air carries operating under I4 CFR Part I35
develop, and include ic their Eight operation manbals and mining
programs, stabilized approact-. criteria. The criteria shou!d include
specific limits of localizer, giideslope, and VOR needle derlections
i5 See Aircraft Accident Report--"iiSAir, kc., night 5050, Boeing 737-4GQ, N416US. Flushing,
New York, September 20, 1989" {NTSBi. iAR-90/03)
'?See Aviation Accident Report--"Cont.-olled Cdlision With Terrain, GP Exp-ess Airlimes, Inc.,
Flight 861, A Beechcraft C 9, N11 SGP, AMiStOn, Alabama, June 8, 1992" (NTSFSB/AAR-93/03)

In a letter dated Jum IC,, 1953, &e FAA advised that it would issue m
ACOB emphasizing stabilized approach criteria information and associated 'uaining
issues, and referencing guidance makiai currently available on &is subject. Bzsed
on this information, oc November 19, 1991. the Safety Board classified A-93-36
"Open--Acceptable Alternate Xesponse."
The Safety Board cmnot understand why the FAA has not yet
completed these actions and issued the applisable ACOB. la any event, the Safety
Board now believes that the AC@B route to address this issue is aot appropria\e. E
a stabilized approach procedure had ken developed and required to be adhered to
by all pilots for night IMC approaches, perhaps this accident would nave been
prevented. Therefore, the Safety I3oa-U classifies A-93-36 "Open--Unacceptable
Response" and reiterates A-93-35. Further, the Safety Board urges the FAA 10
review its position on the need for regulatory action and to mo-e expeditiously
t~ward requiring Part 135 operators to include in theii flight operations manuals and
training programs stabilized approach criteria.
The Safety Board is coficemed that the FAA has not addressed the
passenger and crew safety issue associated with the Pacific Scientific belt design.
The Safety Board reminds the FAA that Title 14 Code of Federal Regulations (CFR)
91-107(a)(3) states, in part, that each person on board a U.S.-registem! civil aircraft
must occupy an approved seat or berth with a safety belt and, if installed, a shoulder
harness h at is properly securcd about him or her during movement on the surface,
takeoff, &xi landing. This same requirement is also reflected in other Yegulations,
such as i4 CFR 135.128(a), and 14 CFR 121.311(b). The Safety Board believes
that if passengers and crew are required by the CFR to wear saiefy belts, then it is
the responsibility of the FAA to ensure that the safety belts function properly.
Although the FAA is in the final stages of issuing an airwo&iness directive to
remove t he safety belts from service, it will take several months tc accomplish this
task. The Safety Board belicves that when passengers board an aircraft, they have
the right ti? ensure that eveqf&iiig on that aircraft is functioning properly. If, as in
this case, the safety belt, under eaergency conditions, may not function as dssigned,
then it is the FAA's responsibility to ensure that operators advise passengers and

70
crew that they must align the insert with the buckle to ensure that the bu~:lde will
reiease should an emergency evacuation become necessary.
Recent conversations that FAA staff have had with Safety Board staff
show &ha: thc FA4 does not intend to take further action on Safety Recommendation
A-94-67, concerning Pacific Scientific safety beit buckles. Therefore the Safety
Box& now classifies Safety Recommendation A-94-67 "Closed--Unacceptable
Action."
The Safety Board strongly believes that until these restmint systems are
replaced, the FAA shouid immediately notify all operators and require them to
explain to passengers and crewmembers, before each flight, how to release these
safety beits baseci upon ale design deficiency found in this investigation.
The Safety Board is currently conducting a safety study of the
star~dards mnd practices in ihe coniuiutei akke kdttstry. !$eve& 5id issues zre
being addressed in the skdy, including: flightcrew training (including the
availabi!ity and use of flght sirnutatom); flightcrew scheduling ad crew pairiig
policies; crew resource management (CRW training; the certification and design of
commuter airplanes; management oversight; ad FAA smeilimce. ais study was
initiated in the spring of 1994, aqd the fiid report is scheduled to be prese~ted to
the Board in November 1994.

1.
2.
3,
4.
5.
6.
7.
8.
me airplane was certified, quipped, and mintaind in
accordance with Federal Aviation Regulations and approved
procedures.
The fiightcrew was trained and certified for the €light m
accordance with company procedures a;,5 Federal Aviation
Regulations.
There was no evidence of failures of my of the stmctures,
systems, or engines that contributed to the accident.
The weather was essentially the same as forecast by the National
Weather Service, and the piiots were aware of the current
weather conditions.
Light to moderate mixed icing condiiions existed dmzg the
approach to Columbus; however, airframe icing was not a factor
in the cawe of the accident.
Air tmfflc services were not totally in accordance with
established procedures but did not contribute to the cause of the
accident.
Tie 3-4101 was a new airplane placed into service in the United
Stztes by ACA in May 1993. Both pilots had low fli&t time
and experience in the airplane and in my airplane equipped with
an electronic eight instrument system (ERS). AddiFionally, the
captain had low time and experience as a captain.
High speed approaches to the final approach fi were often
fiou% by 1-4101 Crews, aithough the FiOcedure was neither
pcblished in the company operations and training manuals nor
appmved by the FAA.

(5) The cornpmqr’s failure to provide adequate crew resource
management training, and the FAA’s failure to require such training;
and
:6) The unavailability of suitable training simulators that precluded
fully effective flightcrew trainhg.

74
--to the Federal Aviatbn Ahinis%ation:
Ensure that all Part 135 operam-s that hco~orate both a high sped
approach profie and a coupled approach profile in the trahing
mual for ail airplmes train pilots to proficiency for Lhose
approach profiies. (Class H, Pri~sity Action) (A-94-1?4]

75
Soh K. Ezuber
Member

1. Investigation
77
APPENDIX A
The National Transportation Safety Board was notified of the accident
about 0030 eastern standard time on January 8, 1994. An investigative team was
dispatched from Washington, D. C., early that moming. Yt was composed of the
f~pl~wing pups: operations: air m%c control; weather; structures; systems;
powerplants; survival factors; and aircmft performance. In addition, specialist
reprt were prepared for the CVR, FDR, and human performance.
Parties to the field investigation were the FAA, the National Air Traffic
Controlfers Association, Jetstream Aircraft Limited, Atlantic Coast Airlines, the
Association of Right Attendants, the Air Line Pilots Association, McCauley
Accessory Division, and Allied Signal Corporation. The Air Accidents
Investigation Bmch (aAzB) of the U.K. was notified of the zccident and was
e mntecf stabs in $his investigation in accordance with Ap_.;ex 13 to the Convention
on international Civil Aviation.
A public hearing was not held regarding the accident.

1. Investigation
77
APPENDIX A
The National Transportation Safety Board was notified of the accident
about 0030 eastern standard time on January 8, 1994. An investigative team was
dispatched from Washington, D. C., early that moming. Yt was composed of the
f~pl~wing pups: operations: air m%c control; weather; structures; systems;
powerplants; survival factors; and aircmft performance. In addition, specialist
reprt were prepared for the CVR, FDR, and human performance.
Parties to the field investigation were the FAA, the National Air Traffic
Controlfers Association, Jetstream Aircraft Limited, Atlantic Coast Airlines, the
Association of Right Attendants, the Air Line Pilots Association, McCauley
Accessory Division, and Allied Signal Corporation. The Air Accidents
Investigation Bmch (aAzB) of the U.K. was notified of the zccident and was
e mntecf stabs in $his investigation in accordance with Ap_.;ex 13 to the Convention
on international Civil Aviation.
A public hearing was not held regarding the accident.

Legend of communication descriptions, abbreviations, acronyms and symbols used in the
attached CVR transcript:
CAM
INT
-7
-2
-?
CLE
IND
CMH
TWR
COM
OPS
PA
GPWS
.
#
.,.
0
I]
-
Cockpit area microphone
lntra-cockpit intercom system
Voice for position) identified as Captain
Voice (or position) identified as First Officer
Unidentifiable voice
Cleveland Air Route Traffic Control Center
Indianapolis Air Route Traffic Control Center
Columbus Approach Control
Columbus Tower Local Control
Radio transmissions received by accident aircraft from sources other than those
specificaily listed herein.
Columbus Company Oprations
Aircraft public address system
Ground Proximity Warning System
Uninteiligible word
Expletive deleted
Pause
Questionable text
Editorial insertion
Break in continuity

TIME a
&Q!w!&x wm
225059
CAM (Start of Recording]
2251:05
INT-2 two nlner niner four ... should I quwtion that?
225192
INT-1 a's okajr.
2252:29
iNT-2 yeah we just came in it right on top.
2252:37
INT-1 YW.
225721
INT-2 you got one I'm gonna try two again ... * miles out.
22572.1
INT-1 okay.

2257:31
CMH [the following ATIS repeats three times] port columbus
international airport inlormation alpha time zero three five
zero zulu weather measured ceiling one thousand one
hundred overcast visibility six light snow fog temperature
two three dew point two two winds three three zero at
fOU7 altimeter two niner niner seven ILS runway two eight
left approach in use also landing runway two eight right
.. ail departing aircraft contact clearance delivery one
two six point three prior to taxing ... notice to airmen
taxiway goii two eight left hold short sign out of service
... taxiway bravo hold short sign out of service .. bravo
lour .. advise on initial contact you have iriformation
alpha.
225851
CLE blue ridge two ninetyune contact indianapolis center one
two four point four five.
2260:55
RDO-1 indianapolis center twenty-four forty-five blue ridge two
ninety-one.
2259: 19
RDO-1 indianapolis center blue ridge two ninety-one's with you
at one four thousand.
225923
IMD blue ridge two ninety-one indianapolis center ropr the
altimeter at columbus is two niner niner seven.

2259:20
RDO-I two niner nlner seven blue ridge two ninety-one.
225931
IND blue ridge two ninety-one be advised ah just had a report
of some icing at one four thousand .. ten o'clock to you
and ah about twenty-five thirty miles.
225942
RDO-1 you said that was some light rime? w
2259:44
IND lour zero four charlie kilo what kind of Icing were you
getting?
225947
4Z4CK moderate moderate rime on up to fourteen thousand and
we're ah we're in the clear ah in the clear above us up
here at fifteen thousand.
225956
IND zero lour charlie kilo thank you .. blue ridge two
ninety-one he said it was moderate rime icing up to one
four thousand.
2300:02
RDO-1 okay thanks ah we'll keep that in mind.
2300:06
4Z4CK and we're sitt'n here negative twenty on the centigrade
ah at fifteen thousand for four charlie kilo.
h

3 0
c
0

230254
INT-1 since we're gonna have to k) in this Wie probably end
up get'n ah -
2303:lO
CAM (sound simlar to that of altitude or gnar warning alerl]
2303:1 I
l&T-l thousand.
2303:20
INT-2 we'li probabiy erxi up getting the what?
230322
INT-1
SLRCG we gotta descend down in if rdher then get ii up in
!he clear and keep her as dry as long as pssihte.
230228
RW-1 and indian-lis canter bfue aidge two nlnety-one can we
g0t eh one live thousand for a IItM whila?
2302:33
IND blue ridge two ninety-one stand by.
2302:07
IND blue &!go two ninetyonc climb and maintain one five
thousand.
230250
ROO-1 one live thousand btue, ridge tao nlnntyone.

TIME &
?iQ!mz
2303:32
INT-I
2303:42
INT-2
230353
INT-2
2303:57
INT-1
2304: 17
IN?-l
2394:24
IN r -2
2304:24
INT-1
230426
IFIT-2
2304:29
IldT-1
COCKPIT GO-
CONTRNT
like so ... (then) get pilot's discretion we'll just .. live
hundred to go .. bring her down real quick.
live to go.
traflic nine o'clock ten.
roger.
okay see if you can reach company tell 'em you're i 3bout
eighteen out .. I got one.
alright -
what's the AfIS called agrrin?
columbus oh ATlS is alpha
nlright I got one.
2304:45
RDO-2 columbus ops blue ridge two ninety-one.

2305: 10
CAM [sound of single chime]
2306:t 1
INT-1 okay tell you what don't worry about it ..
230506
RDO-2 and Columbus ops blue ridge two ninety-one.
2305:33
IND blue ridge two ninety-one turn twenty degrees to the left
this is vw.::-!s lor runway two eight at Columbus.
230535
RDO-2 columbus ops blue ridge two ninety-one is trying to reach
you.
2305:38
RDO-1 twenty degrees to the left lor vectors for runway two
eight columbus.
230542
IND blue ridge two ninety-one pilot's discretion maintain one
one thousand,
2305:46
RDO-1 pilot's discretion to one one thousand bluE ridge two
ninety-one.
2308:Oi
RDO-2 and Columbus 09s blue ridge two ninety-one.
w)
VI

TIME &
SOuBa
2306:15
INT-2 can't get nobody
2306: 16
INT-1
230638
INT-2 okay.
uhm ... call in the back .. tell tha folks oh we'll be about
twenty minutes we're twenty minutes out descending just
give a rough estimata of the weather .. ali you need to do
is tell them it's light snow overcast and whd the winds
are and the temperature don't go into any detail.
2306:39
iNT-1 alright . and I got one.
230655
PA-2 and ah ladies and gentlemen ah we'll be stafling our
initial descent into Columbus ah real shortly ah we should
be on the ground in approximately ah eighteen to twenty
minutes .. ah local weather it's ah twenty-three degrees
ah light snow and ah winds ah seem to be coming out of
the ah northwest at four knots and ah we'd iike to ask
cur flight attendant to prepars the cabin lor landing thank
you.
3307:27
INT-2 back up with you.

TIME &
W:&
2307:30
INT-1
230752
INT-1
230840
INT-2
2300:40
INT-1
2308:45
INT-2
2308:56
INT-1
230858
INT-2
okay .. we'll do the iLS for two eight left .. I don't care.
okay I got a April twenty-fourth nineteen ninety-two
eleven one .. ILS to colunibus two eight left two
seventy-nine inbound heading one oh eight point seven
is the ah Ioc frequency three ninety-one is SUMIF which
is the outer marker .. we ah thousand fourteen and a halt
mile we have that ,.. glide slope is ah .. intercept is ah
twenty-seven hundred feet .. missed approach is climb to
twenty-seven hundred leet direct to ah looks like GRENS
locator outer marker and hold .. looks like it's gonna be
right turns .. any questions?
no quzstions.
okay .. ah let's do a descent and approach check.
roger ., I3 and A .. pressurization checked.
checked.
APR is armed.

c
2
a
._
0

TIME &
.ELu.&E
2309:41
IN7--1
2309:43
INT-2
2303:44
INT-1
230958
I wr -2
231O:Ol
INT-1
continue with the checklist.
crew briel.
okay we wanted it ah two eights flaps twenty-five
standard calls ref speeds we might do a AOA depending
or1 what hnppons n3 we won't do an AOA it'll be ah
without AOA so what are the speeds?
okay ref aped is gonna be one oh five.
fivo six and twenty
230928
ADO-1 one one niner nine five blue ridge Wo ninety-orlo.
2309:31
IND sir it's one one niner point one five nineteen fifteen.
2309:34
RDO-1 okay nineteen fifteen ah for blue ridge ah two ninety-onc
thanks.
2309:4O
COM [sound of frequency change tone]
00
Q

TIME 8
f3.Qumz
2310:03
I NT-2
2310:03
INT-1
2310:05
INT-2
2310:06
INT-1
2310:07
INT-2
2310:08
INT-1
2310:08
INT-2
2310:li
INT-1
2310:14
INT-2
yeah.
alright
okay -
any questions?
no.
alright.
descent and approach check is completed.
alright .. I'm gonna talk to him you try and reach
company okay.
roger.

2311~8
1Nr-l you ever get 'em.
2310:16
RDO-1
2310:22
CMH
2310:31
RDO-1
2310:34
RDO-2
2310:46
RDO-2
231056
RDO-2
231 I :oa
RDO-2
columbus approach blue ridge two ninety-one is with you
out of thirteen thousand two hundred for one one
thousand alpha.
blue ridge two ninety-one roger ah looks like heading two
eight five intercept the two eight left localizer maintain
one zero thousand.
two eight five for the intercept for the two eight left
localizer and tllat's ah down to one zero thousand for
blue ridge two ninety-one.
and columbus ops blue ridge two ninety-one.
columbus ops ah blue ridge two nine one.
and columbus ops blue ridge two ninety-one.
and ops blue ridge two ninaty-one is trying
a
c

TIME a
s.QM3.E Lxx!ma
2311:iQ
INT-2 no no
231 121
INT-1 okay screw 'em
231 1:26
INT:; okay we're going .. down to ten thousand 0 and A's been
completed ah .. the on:! thing we have left is reach
company.
23.1 1:45
INT-1 depending upon what we go through I might have you
pop the boots at the outer marker we'll See.
231 1:49
INT-2 okay .. all I'd have to do is hit auto-cycle light up right up
here?
231 155
INT-1 yeah just hit autocycle.
231 1:56
INT-2 right okay
2312:12
INT-2 you got six miles in eleven wndred s? typical -
2312:18
INT-1 oh yeah not worried about that.

2312:35
INS-1
231239
INT-2
231241
INT-1
231329
INT-1
2313:32
INT-2
2313:35
INT-1
2314:12
INT-1
231413
iNT-2
down to four.
that's what I said to him.
yeah I'm just repeating it I heard you.
what's the winds the surface winds down there .. again?
ah three three zero at four knots.
thanks.
tell ysu what.
yeah.
231226
CMW blue ridge two ninety-one descend and maintain four
thousand.
2312:20
RDO-2 ah down to four thousand blue ridge btdo nineiy-one.

z e
>
a,
C
c
._
c
0

2316:ZO
INT-1 what's the ice AOA ah .
2318:26
INT-1 what's the V speed.
231643
CAM [sound of single chime]
231v:46
INT-1 we're gonna do (laps twenty-live ice AOA on so what's
the refspeod lor that _., at this weight?
2317:19
INT-1 ?housand.
zo1r:m
INS-2 ref's gonna be -
231628
CMH blue ridge two ninety-one is one zero miles from SUMlE
malntaln three thousand until establish on the localizer
cleared the ILS runway two eight lalt approach.
2316:36
RDO-2 ah roger maintain three thousand until established and
ah cleared for the ILS ah two eight left blue ridge two
niner one.
8

TIME &
s!lNBsx CONTENT
2317:20
CAM [sound similar to that of altitude of gear warning alert]
2317:21
INT.2 one to go ... ref ,s gonna be one twelve
2317:25
IN'T-1 what that's VJith the ice AOA right?
2317:28
INT-2 that's affirm,
2317:29
wr-1 okay that's what we're gonna do ... that's what We're
gonna do
231?:46
C A h.1 [soucd of single chime]
2317:5R
CAM [sound similar to reductim in prop/engine noise
anlplitude]
TIME a
sQ!lrsE
2317:43
CMH blue ridge two ninety-one reduce speed to one seven
zero contact tower one three two point seven.
2317:49
RDO-2 one three two point seven on the frequency and reduce
speed to one seventy blue ridge ah two ninety-one.

TIME 8
s!at!BGE
2018:26
INT-1 two ninzty-one.
2310:36
INT-2 what did I say?
2318:38
INT-1 three ninety-one.
2318:39
INT-2 oh.
2318:40
INT-1 okay if you got all the speeds don't worry about them
anymore.
2318:13
COM [sound of frequency change tone]
231890
RDO-2 ah good evening tower blue ridge three ninety-one is with
you on the localizer for ah two eight left.
2318:27
TWR blue ridge two ninety-one columbus tower runway two
eight left cleared to land wind three zero zero at four.
231833
RDO-2 cleared to land blue ridge two ninmy-one.
\B
-4

TIME a
s w
2318:44
INT-2
2318:46
INT-1
231853
INT-2
2319:14
CAM
2319:22
INT-1
2319:30
INT-I
2319:32.0
IFJT-2
2319:36.8
INT-1
2319:39.7
INT-2
2320:01.3
INT-2
INTRA-COCKPIT COMM!.!NICATIO$
ref is one twelve I gotta plug that (too)
I did it !or you.
here comes glide slope.
[sound similar to altitude or gear warning alert]
gimme another one of those.
and we're marker inbound
roger.
don't forget to give me my calls .. a thousand fourteen is
DH.
a thousand .. okay
a thousand above.

COCKPIT CO~UNlCATlOM
TIME a
SOUHCE ?zamim
2320:02.3
INT-1 okay .. flaps nine.
2320:08.5
INT-I gear down.
2320 10.5
CAM [sound similar to landing gear extension]
2320:13.1
INT-2 flaps nine ., waiting for three green.
2329:20.0
iNT-1 flaps fifteen landing chocks.
2320:25.6
INT-7 flaps fifteen landing gear down three green.
2320:28.4
INT-1 landing gear down three green ‘laps liftaeri set indicating.
2320:31.6
INT-2 condition levers .. a hund- conditicn levers a hundred
percent.
2320:36.1
INT-1 okay give me a hundred percent please.
232G:38.1
INT-2 a hundred percent .. flows at lhtee.

2320:39.8
CAM [sound 01 increase in proplengine rpm]
2320:4 1 . 1
INT-I three.
2320:41.6
INT-2 yaw damper.
2320:42.7
INT-1 and autopilot to QO .. don't touch.
2320:44.5
INT-2 don't touch
23%0:46.2
INT-2 lholding on the yaw darnper
2320:46.6
CAM [sound similar to that 01 stick shaker ~idrts]
23%0:47.;!
CAM [sound of seven tones similar to that ol autopilot
disconnect alert]
2320:40.1
INT-1 tony.
2320:49.5
CAM [sound similar to that of stick shaker stops]

232057.5
CAM [sound similar to that of change In or addition to stick
shaker]
232058.7
iNT-1 whoa.
2321 :00.2
CAM [notmd of impact)
2321:oo.a
CAM [End of Recording]

Offre of the Chairman
103
AIPPEMDBX C
SAFETY RECOMMENDATIONS
Honorable David R. Hinson
Administrator
Federal Aviation ndninistration
Yashinqton, D.C. 20591
Dear Vz. Hinson:
Thank you €or the Federal Aviation Administration’s (FAA)
response of May 25, 1992, to the National Trar,sportation Safety
Board’s Safety Recommendations A-94-70 through -72.
Safeky Reconmendation A-94-70 asked the FAA to conduct an
in-depth review of its policies and procedures for the processing
of Air Carrier Operations Bulletins (ACOB), and develop a system to
ensure that the safety information contained therein is acted on in
a timely and accurate manner. The system should include a process
to verify that the actions conteaplated by the ACOB are effectively
implemented.
The Safety Board notes that the FAA will issue a handbook
bulletin to establish a prccess by which all flight s%andards field
offices will acconplish and document surveillance, inspection, or
certificate management actions required by ACOBs, flight standards
informati.on bulletins, and handbook bulletins. Provided this
handbook bulletin is issued to all FSDOs in a timely manner, the
safety Board classifies safety Recommendation A-94-70
“Open--Acceptable Response.” Additionally, the safety Board
requests a copy of the handbook bulletin when it is issued.
Safety Reconmendation X-94-71 asked the FAA to issue immediate
guidance to all principal operations inspectors (POIs) to verify
that the intended safety-related actions contained in ACOB 8-93-4
have been accomplished for air carriers under their jurisdiction.
The Safety Board notes that the FAA will issue a notice
directing its POKs to verify that the actions Contained in ACOB
8-93-4 have been accomplished for the air carriers under their
jurisdiction. The Safety Board classifies Safety Recommendation
A-94-71 “Open--Acceptable RESpGnSe“ snd awaits a copy of the
subject bulletin.
Safety Recoxsendation A-54-72 asked the FL4 to take the
appropriate 2ctior.s to verify that ACOLis issued in the past few
years have Seen inpienente8 as intended.

The Safety Board notes that the FAA will issue a notice
directing its FSDO managers to verify that the actions contained in
all ACOBs issued since January 1, 1992, have been accomplished.
Fending the issuance of the notice, the Safety Boar6 classifies
Safety Recommendation A-94-72 "Open--Acceptable Response.''
Sincerely,
mgina 8igmd 3Y
James E. Xiell
3is Hall
Acting Chairman
cc: ~r.
Donalc? R. Trilling.
Director
Office of Transportation Regulatory Affairs

The Hor:orable Carl W. Vogt
~haiman, National Transportation
-
Safety Board
490 L'Enfant Plaza East. SW.
Uear itr. Chaiman:
This is in response c to Safety Recomendations.A-94-70 through
- -72-issued by the Board on March 17, 1994. These safety
recox=er,datians were issued as a result of the Board's concern
rqarding the.process for disseninatinq air carrier operations
t;&letins (~~02:.
&-9:-70. conduct an in-kp!-fi review of its pclicles and
prcce?;res for the processing.of ACOBs, and develop a systen to
enscre that the safety informatien contained therein is acted
on in a tinely and accurate manner. Tho system should include
a process to verify that the actions contemplated by the kCOB
are effectively inplenented.
- F'S-2. CcT-ent. The Federal Aviation Administration (FAA) will
iss.;e a handbook bulletin to establish a process by which all
flicht standards fieid offices must accolnplish and document
scrveillance, i-spection, or certificate management actions
reyJire6 by ACCB's, flight standards information bulletins, and
ha-dbsck hlletins. This bulletin will direct e2ch flight
szandards district field office nanaqer to lnaintain a master
ccyy of all policy bulletins. The manager will sign each
bulletin upon receipt and'ensure that the appropriate
irspctors receive the bulletins.
7s ezsilre *a-
L.iat this action is being accbnplished, eacn
izs;ectcz k'ho has the required action must make an entry into
tke ;.roSraz Trackincj an2 Reporting S'ubsysten [PTRS) under
FZ.5 CoSes 13E1, 3381, or 5381 to ind!cate that he/she has
ccntzcted the appropriate air carrier.
As a icrther €0~10'mp, regional flight standards divisions will
Grsviee a report of each field office compliance with ACOE's,
fllc5.t standards inforxtion bulletins, and handbook bulletins.

106
This report will be submitted biannually io the Flight
Standards National Field Office.'
I have enclosed a draft copy of the balletin fcr the Board's
infarnation. I will provide the Board with a copy of the final
bulletin as soon as it is issued.
A-94-71. Issue imediate cgiidance to all POIS to verify that
tke ir,tenc?ed safety-related actions contained in ACGB 8-93-4
have been accomplished for air carriers under their
jurisdiction.
FPA-Cb-ment. The FAA will issue n'noticc directing its
principal operations inspectors to verify that the actions
contained-in ACOB 8-93-4,have been a==omplished tor the cir
carriers unaer their jurisdiction. X t is dnticipated that the
notice will be issued bv Yl.21~ 31.
I vi12 provide the Board with a copy oi tne xaoelce as soon as
it is issued.
$.-9c-72. Take the appropriate actions to verl5y that ACOB5
issued in the past few vears have~.been implemented as intended.
p3. Co?xent. The FAA will ~ssue a notice directing its flight
star,dards field office managers'to verify tbat'tha actions
concained in all ACOB's issued since January 1, 1992, have been
accozplished. It is anticipated that this notice will be
issued by JU~Y 31.
I vi21 provide the Board with a copy of the notice as soon as
it 6s issaea.
Sincerely,
Enclosure

Honorable David R. Hinson
Administrator
Federai Aviation Administration
Washington, D.C. 20591
National ~ ~ ~ ~ ~ ~ Safety ~ r 6swd t a t ~ Q n
Washington, D.C. 20594
Safety Recommendation
Date: March 17, 1994
In reply refer to: A-94-71) through -
72
On November 21, 1991, as the result of the investigation of two commuter
airline accidents,' the National Transportation Safety Board adopted Safety
Recommezdation A-91-122, which urged the Federal Aviation Administration
(FAA) to:
Issue an Operations Bulletin to the Principal Operations Inspectors (POk)
of 14 Code of Federal Regulations (CFR) 121 and Pari 135 air carriers to
verify that air carriers have established procedures for flightcrews to take
appropriate actions when they have encountered icing conditions during a
flight, to check fGr the presence of, and to rid airplanes of accumulated
airframe ice prior to ~tiating final approach, in accordance with the
airpiax manufacturers' recommendations on theuse of d-ice sysremns.
AIso as the result of the Livestigation of the same two accidents, on July 22,
1992, the Safety Board adopted Safety Recommendations A-92-59, -60, and -61,
which urged the FAA. to:
'WA inc.. d/b/a Unircd Express. tlig!,f 2415. a Brilish Aerospoce BAJIOI Jelrlrwm, N41OilE. Tri-
ti:ics Airport. Psco. Wasiiinglcn. Wember 25. 1989 (Xi-TSB/AAR-91,G6): and CC Air aritish Acrosp:re BA-
310; Jc~wczrn. XI47PC. Beckley. West Virginia, Jm131)'20. 1991.
6297

i ^
i ? 3
~

equipment. The design changes to the safety be;t tmckles Were
deve?.cped by the manufacturer, in coopexation with Federal
Xviaticn Ad-inistration (FA&) engineers and the Civil
Aercaedical Institute. The manufacturer 4s aggressively
pursuing the replacenent of these safety belts. The FAA is
considering the issi-ance of an airvrthiness directive to
require nar,datory repiaceaent of the buckles within 90 days.
If the FAA issues an airdor'ihiness directive it will be sent to
all operators of affected aircraft. In the meantime, the FAA
believes the manufacturers notification to all operators is
sufficient interim action.
- ... -.
z provide the Board with a copy of any docunent that may
be issued.
A-52-65. Anend TSO-CZ2f to incorporate procedures which would
olace ;saterial. representative of soft abiominal tissile between
the test apparatus and the release buckle ts ensure that safety
belts can be released when subjected to loads specified in the
TSO .
€>A C~znerlt. Tetknicai Standard oraer (TSO) C22f was revised
OVQ~ a year agc to address *&e corrcerns expressed in this
recoraendation. Currently, TSO-C22g incorporates Sy reference
the seat belt requirenents of %\e Society of AutoIEotive
Erigineers Aerospace SEandard (AS) &043. The body block used in
AS 8043 prcvides closed. cell nonresilient foam representative
of soft tissue in the area of the seat belt. As reqzired by
14 cFZ 21.605, all ne4 seat beit applicants must comply with
tke requirelenrs of TSc-CZZg.
I consider rho FAA's aczion to be cocpleted 011 this safety
recomaexdatioz, and I plar. no ;,.,er C*-rh action.
Sixcerely,
-

Honorable David R. Hinson
Administrator
Federal Aviation Administration
Washington. D.C. 20593
115
National Transportation Safety Board
Washington, D.C. 20594
Safety Recommendation
Date: mrch 14, 1994
In reply refer to: A-94-67 through -69
On January 7, 1994.2 letstream J4101, N304UE, operated by Atlantic Coast Airlines of
Sterling. Virginiz, as United Express ffigh: 6291, was on a scheduled cornlfiuter flight from
Dulles Internationa! Airport to Port Columbus International Airport, in Gahanna, Ohio. At 2321
=?ern smndard time. while on an instrurnent )anding system (IS! approach to runway 28L, L+e
alrpjane struck 2 concrete block building bar was abouc 1.2 miles east of the runway. The pilot,
co-pi!ot, fligh: attendant, and two passengers were fatally injured, and the three other passengers,
a husband and wife and their 5-year-old daughter, sustained minor injurls The airplane was
d-zs!royed by postcrash fire.
On January 8. !994, the Safety Board interviewed the husband, who is a frequent air
traveler. He srated that his family war. originaliy assigned to seats 3A. 3B, and 3C, but due to
the iight passenger load, for weight and balance purposes he was moved to seat 8B, his daughter
to 8C. and his wife to 7C. Two other male passengers occupied s ~ t6B s and 9B.
The husband stated that the seatbelt and no smoking signs were illuminated for the entire
Sight. 4t abou: 2310. the airplane began descending, and the pilot announced the descent for
landisg. The Ian6ing gear was louered about 5 minutes before the accident. The husband said
thh: .e airplane con:inu& to descend, and that he could see lights on the ground. Suddenly, the
aiiglane rolled about 45" in one direction and then about 45" in the othcr direction--he coilld not
recall uhrcher the first roll was ID the lefi or to the right, only that it happened very quickly.
Af!rr the roll excursions. the husband sated that the airplane was "wobbly' 2nd rhen dropped
for abos I second and stopped. He described the recovery from the airplane's drop as 'cushy,"
then mcments later :he airplane struck the ground.
b.fter the airpiane came to rest. there were no lights in the cabin. 2nd the only
i;?un?ir.ation came fro% a fire in ille lek engine. The husband said that he remaimd in his seat
iipit-fht and that :he seas remained atached IO the airplane's floor. Hcwevei. he said that hz
2xpe::enszc) 2 'terribly difficult time rcmosing his searbelt.- He said that the plastic relwe lever
on the bcckle uas 'difficult" to operare bemuse he believed tha: it had to be moved greater rh2n
E303

egardless of how far the re!ease levers were opened. Two specific conditions were idexified that
prevented the release. The firs: was the geometric relationship of the fiat piate and the "1)"shaped
ho!e in the insert haif, and 'J;e 'D"-Shaped protrusion and the lockbar on file buckle half.
It was found :hat under some circumstances even with the lockbar rotated into the 'release"
position, the end of the flat plate on the insert half would contact the iockbv shaft so that the
insert would not lift complesely off the 'D'-shaped protrusion. This would happen when the
huc;i!eiinsert assembly was subjected to an Gutward load, causing a misalignment between the
iwo pms. Wi01 the release lever held in the normal ielwe position. the insert could be
disengaged from the buckle if pulled outward to align the two parts. The seami condition that
prevented release was when 15,e release lever was pulM past its normal release position to its
fti!: posl:ion. in :his me. the end of the ref-&e lever itself interfered with the end of the
imx! 2nd p:evented the irisert from being raised above the "1)'-shaped pronusion on the bottom
pla!e of :he buck!e. T>is occurred regardless of the alignmen: of the buckle and insert.
On February 8 2nd 9. i594. !he Safety Board and representatives from the Federal
Avialio21 Adrninis:rziion's !FAA) Aircraft Certification Managemefit Office, ierstream Aircraft
Con?pany. A;iantic Coast Airlines, and the Ai; Line Pilors Associatioa met !0 examine the safety
bel! rsleasc buckles ai the Pacific Scien!ific Faciiity. During this meeting, Pacific Scientific
di-r;ons!rzted that !he saiery beirs and release buckles me! the requirements contained in FAA's
TX~C22f. This derno;.s:ration consisted of a passenger safety belt placed around a body block,
a:;L Suckicd, and then loaded irk accordance with the TSO. Once it was demonmated that the
safii:; bcit complied wiin the TSO, a i-inch piece of dense foam was placed between the body
Xtd znd :he si;fe;y bel: to represent the seat occupant': soft abdominal tissue. it was found hat
.> ..i:,. :?
XJLC b~-. c 2p;irr~:rt.zreiy 25,rxO of the passenger acd crewmember restrain: systems of this
Lhig I:i Ljsc ~AS>:l&&.

The Safety Board bei!eves that a11 cperators that use these passeager and crewmember
restraint systems should be no?ified of the Safety 30ard's findings, 2nd that the FAA sbould take
action to require the removal of this design and replacement with resc.air.ts of a different design
as expeditiously as possible consistent with :he avaikbility of replacement buckles. The Safety
Board also believes that until thcse restraint system are ieplaced, he FAA should notify all
operators to inform passengers and crewmembers on how to rslease their safety belts based upon
the design deficiency foilnd in this investigation.
Therefore, based on the above informztion. the Safety Board recommends that the Federa:
P.>:iation Administration:
Immediately notify all operators of the Safety Board's finding. including the U.S.
Department of Defense and foreign governments, and require ail operators whose
sircraf! have the bffected Pacific Scientific safety belt buckics to inform
passengers and crewmembers about !he wed to 2lign t?e buckle inserr to assure
easy release of the safety belts. (Class 1, Urgent Action) (A-94-67)
issue an Airworthiness Directive to reqdi:e the removal and replacement 3f all
safe!); bel& manufactured by Pacific Scientific fo: Pari Number 1108435 buckles,
with the 45'1ift levers, and Put Number 1108460 buckles with the 90" lift levers,
with belts having bwk!cs of a different design as expeditiously as possible,
consistent with the availability of repiacement buckles. (Class I, Urgent Action)
!A-94-68)
Amend TSO-CXf !o incorporate procedures which would place material
repmenrative of soft abdornirai tissue between the test apparztus and the release
buckle :o ewwe th2t szfery be!& can be relwed when subjected to loads specified
in :he TSC). (Class 11. Priority Action) (A-91-69)
Chairman ':iOGT, Vice Chairman COUGHLIN, and Members LAUBER,
tlAhlMERSCHMIDT. and HALL concvrred in these recommendations.
Chairman